UMAP '19- Proceedings of the 27th ACM Conference on User Modeling, Adaptation and PersonalizationFull Citation in the ACM Digital Library
SESSION: Keynote & Invited Talks
To be effective, modern technological applications should take into account the needs, preferences, capabilities, and limitations of human users. In recent years, this requirement has made more imperative the need to understand in more detail the human cognition and its constraints. Cognitive processes and their underlying neural substrates are traditionally investigated with laboratory studies that yield data of different forms, ranging from accuracy and reaction time data in behavioural experiments to electrophysiological responses and neuro-imaging data in neuroscience studies. But how do such data enable psychologists and other scientists to draw conclusions about cognition? Also, how can the extracted knowledge be exploited for the design of evidence-based smart systems and innovative technologies? In this talk, I will address these questions by drawing examples from my research that employs various methods and techniques, including behavioural experiments in Virtual Reality, eye-tracking, and physiological recordings. Although most of this research focuses on how people attend, perceive, and memorize spatial information, studies investigating more general cognitive mechanisms (e.g., selective attention and executive functions) will be also presented.
User engagement plays a central role in companies and organisations operating online services. A main challenge is to leverage knowledge about the online interaction of users to understand what engage them short-term and more importantly long-term. Two critical steps of improving user engagement are defining the right metrics and properly optimising for them. A common way that engagement is measured and understood is through the definition and development of metrics of user satisfaction, which can act as proxy of short-term user engagement, mostly at session level. In the context of recommender systems, developing a better understanding of how users interact (implicit signals) with them during their online session is important for developing metrics of user satisfaction. Detecting and understanding implicit signals of user satisfaction are essential for enhancing the quality of the recommendations. When users interact with the recommendations served to them, they leave behind fine-grained traces of interaction patterns, which can be leveraged to predict how satisfying their experience was. This talk will present various works and personal thoughts on how to measure user engagement. It will discuss the definition and development of metrics of user satisfaction that can be used as proxy of user engagement, and will include cases of good, bad and ugly scenarios. An important message will be to show that, when aiming to personalise the recommendations, it is important to consider the heterogeneity of both user and content to formalise the notion of satisfaction, and in turn design the appropriate satisfaction metrics to capture these.
Researchers claim that we are facing a global loneliness epidemic, and that mental illness, anxiety disorders, stress and burnout are on the rise. Technology, such as social media, is often found to have a detrimental effect on mental health, self-esteem and sleep, and to cause anxiety and feelings of loneliness. This talk is about how adaptive systems can actively improve well-being, instead of contributing to making it worse. We will discuss different ways of doing so, the work already done, the challenges faced, and our vision of a new kind of personalized systems that act as guardian angels. First, systems can provide emotional support, adapted to the recipient's characteristics such as their personality, affective state, cultural background, and stressors experienced. Second, systems can aid humans to provide emotional support. People often struggle to support others, and may say something that is counter productive or nothing at all. Systems can train people on how to provide support. They can also mediate emotional support, adapting support messages to both the support giver and recipient, taking into account for example the closeness of relationships and people's personality. Third, systems can support and motivate people to adopt behaviours that improve their well-being and that of others, and to better regulate their emotions. There has been much research on persuasive technology to support people in changing behaviours, and it has been shown that both the behaviour change techniques used, and attributes of techniques need adapting. Whilst much persuasive technology research has focused on physical well-being and sustainability, the emphasis in this presentation will be on mental well-being and encouraging people to help each other. Fourth, systems can team people up. Systems can decide who are best placed to provide support and motivation, encouraging particular people to support (or ask help from) particular other people. Additionally, adaptive group formation (or peer-to-peer recommendations) can be used for joint problem solving scenarios, with a system deciding or recommending who should work with whom. There are many benefits to group work, but it is also often a source of negative emotions. Adaptive group formation can consider affect and personality in addition to expertise, to minimize such negative emotions. Finally, systems can improve the well-being of groups and not just individuals. People's well-being is influenced by the well-being of others in their surroundings, and people's actions impact the well-being of others. Systems can monitor group well-being. They can encourage and support effective group behaviours, for example, by providing feedback on how group members and the group as a whole function. They can support the building of group identity and cohesion. They can support groups in making decisions that are good for group well-being. Overall, we envision adaptive systems as effective and emotionally intelligent contributors in the community, improving the way people interact, and acting like guardian angels.
SESSION: ACM UMAP 2019 Main Track
In this paper we present a methodology to justify the suggestions generated by a recommendation algorithm through the identification of relevant and distinguishing characteristics of the recommended item, automatically extracted by mining users' reviews. Our approach relies on a combination ofnatural language processing and sentiment analysis techniques, and is based on the following steps: (1) a set of users' reviews discussing the recommended item is gathered and analyzed; (2) the distinguishing aspects that characterize the item are extracted and a ranking function is used to identify the most relevant ones; (3) excerpts of the reviews discussing such aspects are extracted and a natural language template is filled in through the aggregation of these sentences. This represents the final output of the algorithm, which is provided to the user as justification of the recommendation she received. In the experimental evaluation, we carried out a user study (N=296, 73.6% male) aiming to investigate the effectiveness of our methodology in two different domains, as movies and books. Results showed that our technique can provide users with rich and satisfying justifications. Moreover, our experiment also showed that the users prefer review-based justifications to other explanation strategies, and this finding further confirmed the effectiveness of the approach.
Explanations help users to better understand why a set of items has been recommended. Compared to single user recommender systems, explanations in group recommender systems have further goals. Examples thereof are fairness which helps to take into account as much as possible group members' preferences and consensus which persuades group members to agree on a decision. This paper proposes different explanation types and investigates which explanation best helps to increase the fairness perception, consensus perception, and satisfaction of group members with regard to group recommendations. We conducted a user study to evaluate the proposed explanations. The results show that explanations which take into account preferences of all or the majority of group members achieve the best results in terms of the mentioned aspects. Moreover, there exist positive correlations among these aspects, i.e., as the perceived fairness (or the perceived consensus) of explanations increases, so does the satisfaction of users with regard to group recommendations. In addition, in the context of repeated decisions, the inclusion of group members' satisfaction from previous decisions in the explanations helps to improve the fairness perception of users with regard to group recommendations.
Evaluating Visual Explanations for Similarity-Based Recommendations: User Perception and Performance
Recommender system helps users to reduce information overload. In recent years, enhancing explainability in recommender systems has drawn more and more attention in the field of Human-Computer Interaction (HCI). However, it is not clear whether a user-preferred explanation interface can maintain the same level of performance while the users are exploring or comparing the recommendations. In this paper, we introduced a participatory process of designing explanation interfaces with multiple explanatory goals for three similarity-based recommendation models. We investigate the relations of user perception and performance with two user studies. In the first study (N=15), we conducted card-sorting and semi-interview to identify the user preferred interfaces. In the second study (N=18), we carry out a performance-focused evaluation of six explanation interfaces. The result suggests that the user-preferred interface may not guarantee the same level of performance.
Structured user model data not only allow system personalization, but also may be of interest as a source for analysis: in particular, for the study of general trends and for the detection of anomalies in preferences and mutually-referenced features among different user models. Such sources are multidimensional and interrelated, and recently started to be represented as graph-based datasets. Among the most effective ways of studying such data is visual exploration based on data-driven graph drawing approaches: in particular, node-link and node-link-group diagrams. The paper provides an overview of advanced approaches to the graphical representation of multidimensional data derived from user modeling and presents a proposal for developing flexible and scalable user interfaces for the hypergraph-based visual exploration of relations within a user model (UM). Then, we propose these principles in the visualization of an existing adaptive system.
Effect of Values and Technology Use on Exercise: Implications for Personalized Behavior Change Interventions
Technology has recently been recruited in the war against the ongoing obesity crisis; however, the adoption of Health & Fitness applications for regular exercise is a struggle. In this study, we present a unique demographically representative dataset of 15k US residents that combines technology use logs with surveys on moral views, human values, and emotional contagion. Combining these data, we provide a holistic view of individuals to model their physical exercise behavior. First, we show which values determine the adoption of Health & Fitness mobile applications, finding that users who prioritize the value of purity and de-emphasize values of conformity, hedonism, and security are more likely to use such apps. Further, we achieve a weighted AUROC of .673 in predicting whether individual exercises, and we also show that the application usage data allows for substantially better classification performance (.608) compared to using basic demographics (.513) or internet browsing data (.546). We also find a strong link of exercise to respondent socioeconomic status, as well as the value of happiness. Using these insights, we propose actionable design guidelines for persuasive technologies targeting health behavior modification.
In recent years, recommender systems have emerged as a key component for personalization in health applications. Central in the development of recommender systems is rating-based preference elicitation, based both on single-criterion and multi-criteria rating. Though its use has already been studied in various domains of recommender systems, far too little attention has been paid to preference elicitation in health recommender systems~(HRS). The purpose of this paper is to develop a better understanding of this preference elicitation by studying the criteria that users consider when they rate a health promotion recommendation from HRS, and accordingly, to offer a design solution as a functional feedback model for mobile health applications. This paper investigates the user-perceived importance of various criteria, as well as latent factors for eliciting user feedback on the recommendations. It also reports the relationship of explanation and trust to the overall rating. By aggregating a list of all possible criteria, we further discover that not all criteria are equally important to users, and that the effectiveness of a recommendation plays a dominant role.
People with Multiple Sclerosis (pwMS) suffer from a diverse set of symptoms such as fatigue, pain, depression, and decline in motor and cognitive function. It has been proven that physical activity has a positive effect on most of these symptoms. However, many pwMS lead sedentary lives, and do not meet the guidelines for physical activity. We propose WalkWithMe, a mobile application that supports pwMS in walking. WalkWithMe coaches pwMS in achieving a personal goal over a period of 10 weeks. We conducted a workshop with pwMS and brainstorm sessions with experts in rehabilitation to define the design choices of WalkWithMe. We examined the impact of WalkWithMe in a 10-week field study with 13 pwMS. The study revealed insights in walking habits, and positive trends in walking capacity. In this paper, we present the design aspects of WalkWithMe, findings of our 10-week evaluation, and resulting insights on goal setting for pwMS.
We present insights obtained from a web-based game designed to investigate trust-related factors in a care-taking scenario. The game is set in a retirement village, where elderly residents live in smart homes equipped with monitoring systems. These systems should raise alerts when adverse events happen, but they do not function perfectly (they may issue false alerts or miss true events). Players, who "work'' in the village, perform a primary task, whereby they must ensure the welfare of the residents by attending to adverse events in a timely manner, and a secondary routine task that demands their attention. Our contributions are (1) the game itself, which supports experimentation with various trust-related factors; (2) a methodology for the calibration of the game's parameters; (3) insights from two experiments regarding the relationship between device performance, in particular error type, and trust and user behaviour; and (4) insights from predictive models about factors that influence trust and aspects of user behaviour.
The Clock Drawing Test is used as a cognitive assessment tool in geriatrics to detect signs of dementia or to model the progress of stroke recovery. The result is scored manually by a trained professional. We implement the Mendez scoring scheme and create a hierarchy of error categories that model the test characteristics of the clock drawing test, based on a set of impaired clock examples provided by a geriatrics clinic. Using a digital pen we recorded 120 clock samples for evaluating the automatic scoring system, with a total of 2400 error samples distributed over the 20 error classes of the Mendez scoring scheme. Error classes are scored automatically using a handwriting and gesture recognition framework. Results show that we provide a clinically relevant cognitive model for each subject. In addition, we heavily reduce the time spent on manual scoring. We compare manual scoring results with results produced by our automated system.
An Unscented Hound for Working Memory (AUHWM) is a new framework for the real-time tracking of human Working Memory (WM) that can be used to adapt computer interfaces to users' available cognitive resources. WM is the part of human cognition responsible for the short term storing and handling of information; it can, in stressful situations, under information overload or when suffering from dementia-like diseases, become severely limited, possibly leading to poor decision making. Our preliminary results suggest that AUHWM can provide a precise and timely assessment of WM capacity, so that the cognitive load a specific task imposes on users can be adapted, e.g., at the User Interface (UI) level. AUHWM is based on a low-level stochastic discrete model of human WM dynamics, implemented as a Gradient-Boosting-derived deterministic algorithm that simulates users' oblivion. AUHWM also performs Unscented Kalman filtering to track users' WM-specific parameters in real time, thus providing a dynamic assessment of their cognitive resources. Our approach has been tested and validated using data collected from Match$ ^2$s, a visual memory game played by 18 users in another study. Going beyond real-time WM tracking, AUHWM is intended to also be used for WM prediction, paving the way to the adaptation of tasks and their UIs in real time as a function of users' cognitive abilities; we detail an example of such an adapted system, and provide experimental evidence this approach could lead to future enhanced WM-adapted UIs.
Combining Trending Scan Paths with Arousal to Model Visual Behaviour on the Web: A Case Study of Neurotypical People vs People with Autism
People with autism often exhibit different visual behaviours from neurotypical users. To explore how these differences are exhibited on the Web, we model visual behaviour by combining pupillary response, which is an unobtrusive measure of physiological arousal, with eye-tracking scan paths that indicate visual attention. We evaluated our approach with two populations: 19 neurotypical users and 19 users with autism. We observe differences in their visual behaviours as, in certain instances, individuals with autism exhibit a lower arousal response to affective contents. While this is consistent with the literature on autism, we confirm this phenomenon on the Web. We discuss how our modelling method can be used to identify possible UX issues such as the presence of stress, cognitive load and differences in the perception of Web elements in relation to physiological arousal.
Image analysis algorithms have been a boon to personalization in digital systems and are now widely available via easy-to-use APIs. However, it is important to ensure that they behave fairly in applications that involve processing images of people, such as dating apps. We conduct an experiment to shed light on the factors influencing the perception of "fairness." Participants are shown a photo along with two descriptions (human- and algorithm-generated). They are then asked to indicate which is "more fair" in the context of a dating site, and explain their reasoning. We vary a number of factors, including the gender, race and attractiveness of the person in the photo. While participants generally found human-generated tags to be more fair, API tags were judged as being more fair in one setting - where the image depicted an "attractive," white individual. In their explanations, participants often mention accuracy, as well as the objectivity/subjectivity of the tags in the description. We relate our work to the ongoing conversation about fairness in opaque tools like image tagging APIs, and their potential to result in harm.
Beggars Can't Be Choosers: Augmenting Sparse Data for Embedding-Based Product Recommendations in Retail Stores
Recommender systems are an essential component in many e-commerce platforms to drive sales and guide customers when exploring new products. With the increasing adoption of RFID technology in traditional brick-and-mortar stores, for example, in the form of smart fitting rooms that allow to display recommendations in the integrated mirror, retailers have only recently started to tap into existing product recommendation algorithms. However, due to limited data availability as well as sparsity, for example due to assortments adapted for different demographics, traditional retailers largely struggle to leverage this technology. In this paper we extend the state-of-the-art embedding-based recommender approach prod2vec by processing information about co-purchased products (i.e., shopping baskets) in retail stores. By adding point-of-sale information to shopping baskets we are able to provide recommendations aimed at individual stores, without having to maintain separate models for each location. Furthermore, we experiment with data augmentation methods to overcome the imposed limitations of the available data, and are able to increase the quality of the computed recommendations by more than 6.9%.
Badges are endemic to online interaction sites, from Question and Answer (Q&A) websites to ride sharing, as systems for rewarding participants for their contributions. This paper studies how badge design affects people's contributions and behavior over time. Past work has shown that badges "steer'' people's behavior toward substantially increasing the amount of contributions before obtaining the badge, and immediately decreasing their contributions thereafter, returning to their baseline contribution levels. In contrast, we find that the steering effect depends on the type of user, as modeled by the rate and intensity of the user's contributions. We use these measures to distinguish between different groups of user activity, including users who are not affected by the badge system despite being significant contributors to the site. We provide a predictive model of how users change their activity group over the course of their lifetime in the system. We demonstrate our approach empirically in three different Q&A sites on Stack Exchange with hundreds of thousands of users, and we discuss the implications for system designers.
Constraint-based group recommender systems support the identification of items that best match the individual preferences of all group members. In cases where the requirements of the group members are inconsistent with the underlying constraint set, the group needs to be supported such that an appropriate solution can be found. In this paper, we present a guided approach that determines socially-aware diagnoses based on different aggregation functions. We analyzed the prediction quality of different aggregation functions by using data collected in a user study. The results indicate that those diagnoses guided by the Least Misery aggregation function achieve a higher prediction quality compared to the Average Voting, Most Pleasure, and Majority Voting. Moreover, another major outcome of our work reveals that diagnoses based on aggregation functions outperform basic approaches such as Breadth First Search and Direct Diagnosis.
Tourist groups exploring a city often face the problem of finding a sequence of points of interest that satisfies all group members. In this work, we present three different configurations of a group recommender system that suggests such trips even when tourists are already traveling: connecting multiple smartphones, sharing a public display, and combining both devices in a distributed user interface approach. We conducted a large user study with real groups to evaluate these configurations. Our results show that public displays are attractive for users who prefer an open discussion of their preferences. However, we have empirical evidence that decisions on group preferences often tend to be unfair for some group members, especially when they do not know each other very well. A distributed recommender system aggregating group members' individual preferences fairly with the option to display selected content on a public display was the most appreciated solution for overcoming this problem.
Social-based recommenders seek to exploit the mechanisms of homophily and influence observed in social networks in order to provide more accurate recommendations. The way they achieve this is by enforcing similar preferences among users that are socially connected. It is thus reasonable to question whether such approaches lead to the formation of echo chambers, i.e., social groups with a narrow set of preferences and which receive recommendations with low diversity and novelty. This work studies this research question and quantifies the diversity and novelty of existing methods. An important finding is that it is possible to increase accuracy without sacrificing diversity and novelty.
Novelty enhancement of recommendations is typically achieved through a post-filtering process applied on a candidate set of items. While it is an effective method, its performance heavily depends on the quality of a baseline algorithm, and many of the state-of-the-art algorithms generate recommendations that are relatively similar to what the user has interacted with in the past. In this paper we explore the use of sampling as a means of novelty enhancement in the Bayesian Personalized Ranking objective. We evaluate the proposed extensions on the MovieLens 20M dataset, and show that the proposed method can be successfully used instead of two-step reranking, as it offers comparable and better accuracy/novelty tradeoffs, and more unique recommendations.
This paper focuses on recommender systems based on item-item collaborative filtering (CF). Although research on item-based methods is not new, current literature does not provide any reliable insight on how to estimate confidence of recommendations. The goal of this paper is to fill this gap, by investigating the conditions under which item-based recommendations will succeed or fail for a specific user. We formalize the item-based CF problem as an eigenvalue problem, where estimated ratings are equivalent to the true (unknown) ratings multiplied by a user-specific eigenvalue of the similarity matrix. We show that the magnitude of the eigenvalue related to a user is proportional to the accuracy of recommendations for that user. We define a confidence parameter called the eigenvalue confidence index, analogous to the eigenvalue of the similarity matrix, but simpler to be computed. We also show how to extend the eigenvalue confidence index to matrix-factorization algorithms. A comprehensive set of experiments on five datasets show that the eigenvalue confidence index is effective in predicting, for each user, the quality of recommendations. On average, our confidence index is 3 times more correlated with MAP with respect to previous confidence estimates.
When evaluating algorithms that recommend a list of relevant items to a user, it is common to use metrics such as precision to measure the system accuracy. When computing precision, one computes the number of items that were selected by the user among the recommended items. As such, recommended items that were not selected by the user, which we call \em rejected recommendations, are all considered to be bad recommendations, resulting in no increase to the system accuracy metric. Our ultimate goal is to develop a new recommendation accuracy evaluation metric, which may assign some value to the rejected recommendations. In this paper, as a first step, we claim that some rejected recommendations are better than others. Specifically, we consider items that are similar to the item that was finally selected, as better recommendations than items that bear little similarity. We conduct a user study, showing that rejected recommendations that have high content or collaborative similarity to the selected item are perceived by users as better recommendations than items with low similarity. In addition, we study the correlations between the recommended items shown to a user and the un-recommended items that the user has selected in a real-life job posting dataset. We show that when considering item similarity rather than simple precision, the correlations are much higher. This may be attributed to the influence of the recommended items on the decisions of the user.
Web Extensions (add-ons) allow clients to customize their Web browsing experience through the addition of auxiliary features to their browsers. The add-on ecosystem is a market differentiator for the Firefox browser, offering contributions from both commercial entities and community developers. In this paper, we present the Telemetry-Aware Add-on Recommender (TAAR), a system for recommending add-ons to Firefox users by leveraging separate models trained to three main sources of user data: the set of add-ons a user already has installed; usage and interaction data (browser Telemetry); and the language setting of the user's browser (locale). We build individual recommendation models for each of these data sources, and combine the recommendations they generate using a linear stacking ensemble method. Our method employs a novel penalty function for tuning weight parameters, which is adapted from the log likelihood ratio cost function, allowing us to scale the penalty of both correct and incorrect recommendations using the confidence weights associated with the individual component model recommendations. This modular approach provides a way to offer relevant personalized recommendations while respecting Firefox's granular privacy preferences and adhering to Mozilla's lean data collection policy. To evaluate our recommender system, we ran a large-scale randomized experiment that was deployed to 350,000 Firefox users and localized to 11 languages. We found that, overall, users were 4.4% more likely to install add-ons recommended by our ensemble method compared to a curated list. Furthermore, the magnitude of the increase varies significantly across locales, achieving over 8% improvement among German-language users.
Interactive adaptive systems powered by Reinforcement Learning (RL) have many potential applications, such as intelligent tutoring systems. In such systems there is typically an external human system designer that is creating, monitoring and modifying the interactive adaptive system, trying to improve its performance on the target outcomes. In this paper we focus on algorithmic foundation of how to help the system designer choose the set of sensors or features to define the observation space used by reinforcement learning agent to make decisions. We present an algorithm, value driven representation (VDR), that can iteratively and adaptively augment the observation space of a reinforcement learning agent so that is sufficient to capture a (near) optimal policy. To do so we introduce a new method to optimistically estimate the value of a policy using offline simulated Monte Carlo rollouts. We evaluate the performance of our approach on standard RL benchmarks with simulated humans and demonstrate significant improvement over prior baselines.
Collaborative Filtering is largely applied to personalize item recommendation but its performance is affected by the sparsity of rating data. In order to address this issue, recent systems have been developed to improve recommendation by extracting latent factors from the rating matrices, or by exploiting trust relations established among users in social networks. In this work, we are interested in evaluating whether other sources of preference information than ratings and social ties can be used to improve recommendation performance. Specifically, we aim at testing whether the integration of frequently co-occurring interests in information search logs can improve recommendation performance in User-to-User Collaborative Filtering (U2UCF). For this purpose, we propose the Extended Category-based Collaborative Filtering (ECCF) recommender, which enriches category-based user profiles derived from the analysis of rating behavior with data categories that are frequently searched together by people in search sessions. We test our model using a big rating dataset and a log of a largely used search engine to extract the co-occurrence of interests. The experiments show that ECCF outperforms U2UCF and category-based collaborative recommendation in accuracy, MRR, diversity of recommendations and user coverage. Moreover, it outperforms the SVD++ Matrix Factorization algorithm in accuracy and diversity of recommendation lists.
Linguistic Design of In-Vehicle Prompts in Adaptive Dialog Systems: An Analysis of Potential Factors Involved in the Perception of Naturalness
Against the background of current trends towards natural and adaptive in-vehicle Spoken Dialog Systems, this paper aims at evaluating potential factors involved in the perception of naturalness and comprehensibility of system prompts. By conducting an exploratory user study investigating various syntactic paraphrases, we were able to identify several system- and user-sided characteristics which should be considered in the design of system prompts. We conclude from our results that the choice of a syntactic structure for in-vehicle prompts is a relevant question and interestingly depends on several individual user characteristics, such as personality.
Voice User Interfaces (VUIs) are becoming increasingly popular. However, how VUIs can adapt to user differences remains insufficiently understood. We analyze usage data from a user study (n=50) where participants interacted with an unfamiliar VUI. Through automated clustering and statistical analysis, we present user models of their behavior patterns. We found user behavior can be grouped into three clusters: people who become proficient with the system and typically stay proficient while completing different tasks, people who exhibit an exploratory approach to completing tasks, and people who struggled to complete tasks. We discuss design implications based on these behavior clusters.
On the Accuracy of Eye Gaze-driven Classifiers for Predicting Image Content Familiarity in Graphical Passwords
Graphical passwords leverage the picture superiority effect to enhance memorability, and reflect today's haptic users' interaction realms. Images related to users' past sociocultural experiences (e.g., retrospective) enable the creation of memorable and secure passwords, while randomly system-assigned images (e.g., generic) lead to easy-to-predict hotspot regions within graphical password schemes. What remains rather unexplored is whether the image type could be inferred during the password creation. In this work, we present a between-subjects user study in which 37 participants completed a recall-based graphical password creation task with retrospective and generic images, while we were capturing their visual behavior. We found that the image type can be inferred within a few seconds in real-time. User adaptive mechanisms might benefit from our work's findings, by providing users early feedback whether they are moving towards the creation of a weak graphical password.
Passive and effortless authentication of the owner of wearable devices can be achieved by building a personalized model of his/her movements during gait periods. In this paper, an authentication method based on the distances between a set of body-worn devices is proposed. The method assumes that no prior information is available about users different from the legitimate one. One-class classification methods are used to distinguish the gait segments of the owner from the gait segments of possible impostors. Experimental results show that accuracy values as high as ~87-91% can be obtained. The impact of different walking styles (normal, fast, slow, and carrying a bag) is also evaluated.
Persuasion is one of the most frequent, albeit challenging, tasks in human interaction. In a textual argument, one party (author) aims to change the view of the other party (reader). In this paper, we propose to detect persuasive textual arguments while considering the parties personality traits. We find that we can substantially improve accuracy by introducing features that capture author-reader personality traits and their interaction. Our model improves performance of state-of-the-art baselines from 66% to 71% on a new dataset of more than 19K arguments we collected.
Many collaborative recommender systems leverage social correlation theories to improve suggestion performance. However, they focus on explicit relations between users and they leave out other types of information that can contribute to determine users' global reputation; e.g., public recognition of reviewers' quality.
We are interested in understanding if and when these additional types of feedback improve Top-N recommendation. For this purpose, we propose a multi-faceted trust model to integrate local trust, represented by social links, with various types of global trust evidence provided by social networks. We aim at identifying general classes of data in order to make our model applicable to different case studies. Then, we test the model by applying it to a variant of User-to-User Collaborative filtering (U2UCF) which supports the fusion of rating similarity, local trust derived from social relations, and multi-faceted reputation for rating prediction.
We test our model on two datasets: the Yelp one publishes generic friend relations between users but provides different types of trust feedback, including user profile endorsements. The LibraryThing dataset offers fewer types of feedback but it provides more selective friend relations aimed at content sharing. The results of our experiments show that, on the Yelp dataset, our model outperforms both U2UCF and state-of-the-art trust-based recommenders that only use rating similarity and social relations. Differently, in the LibraryThing dataset, the combination of social relations and rating similarity achieves the best results. The lesson we learn is that multi-faceted trust can be a valuable type of information for recommendation. However, before using it in an application domain, an analysis of the type and amount of available trust evidence has to be done to assess its real impact on recommendation performance.
Like other social systems, in collaborative filtering a small number of "influential" users may have a large impact on the recommendations of other users, thus affecting the overall behavior of the system. Identifying influential users and studying their impact on other users is an important problem because it provides insight into how small groups can inadvertently or intentionally affect the behavior of the system as a whole. Modeling these influences can also shed light on patterns and relationships that would otherwise be difficult to discern, hopefully leading to more transparency in how the system generates personalized content. In this work we first formalize the notion of "influence" in collaborative filtering using an Influence Discrimination Model. We then empirically identify and characterize influential users and analyze their impact on the system under different underlying recommendation algorithms and across three different recommendation domains: job, movie and book recommendations. Insights from these experiments can help in designing systems that are not only optimized for accuracy, but are also tuned to mitigate the impact of influential users when it might lead to potential imbalance or unfairness in the system's outcomes.
Today more and more people use social networks and so the differences in personalities of users become more diversified. The same holds true for available news content. To test if regular news and fake news are distributed similarly and to what extent this depends on the personality and behavior of individuals, we conducted a mixed-method study. Through an online questionnaire we measured personality traits of individuals in social networks, how they behave, and how they are connected to each other. Using this data, we developed an agent-based model of an online social network. Using our model, an average of 92% of regular news and 98% of fake news were disseminated to the whole network. Network density turned out to be more important for dissemination than the differences in personality and behavior of individuals. Thus the spread of fake news can not only be addressed by focusing on the personality of individual users and their associated behavior. Systemic approaches---integrating both human and algorithm---must be considered to effectively combat fake news.
Conversational interfaces can facilitate human-computer interactions. Whether or not conversational interfaces can improve worker experience and work quality in crowdsourcing marketplaces has remained unanswered. We investigate the suitability of text-based conversational interfaces for microtask crowdsourcing. We designed a rigorous experimental campaign aimed at gauging the interest and acceptance by crowdworkers for this type of work interface. We compared Web and conversational interfaces for five common microtask types and measured the execution time, quality of work, and the perceived satisfaction of 316 workers recruited from the FigureEight platform. We show that conversational interfaces can be used effectively for crowdsourcing microtasks, resulting in a high satisfaction from workers, and without having a negative impact on task execution time or work quality.
Web browsing has never been easy for blind people, primarily due to the serial press-and-listen interaction mode of screen readers -- their "go-to'' assistive technology. Even simple navigational browsing actions on a page require a multitude of shortcuts. Auto-suggesting the next browsing action has the potential to assist blind users in swiftly completing various tasks with minimal effort. Extant auto-suggest feature in web pages is limited to filling form fields; in this paper, we generalize it to any web screen-reading browsing action, e.g., navigation, selection, etc. Towards that, we introduce SuggestOmatic, a personalized and scalable unsupervised approach for predicting the most likely next browsing action of the user, and proactively suggesting it to the user so that the user can avoid pressing a lot of shortcuts to complete that action. SuggestOmatic rests on two key ideas. First, it exploits the user's Action History to identify and suggest a small set of browsing actions that will, with high likelihood, contain an action which the user will want to do next, and the chosen action is executed automatically. Second, the Action History is represented as an abstract temporal sequence of operations over semantic web entities called Logical Segments - a collection of related HTML elements, e.g., widgets, search results, menus, forms, etc.; this semantics-based abstract representation of browsing actions in the Action History makes SuggestOmatic scalable across websites, i.e., actions recorded in one website can be used to make suggestions for other similar websites. We also describe an interface that uses an off-the-shelf physical Dial as an input device that enables SuggestOmatic to work with any screen reader. The results of a user study with 12 blind participants indicate that SuggestOmatic can significantly reduce the browsing task times by as much as 29% when compared with a hand-crafted macro-based web automation solution.
Identifying instances when a user will not able to attend to an incoming message and constructing an auto-response with relevant contextual information may help reduce social pressures to immediately respond that many users face. Mobile messaging behavior often varies from one person to another. As a result, compared to a generic model considering profiles of several users, a personalized model can capture a user's messaging behavior more accurately to predict their inattentive states. However, creating accurate personalized models requires a non-trivial amount of individual data, which is often not available for new users. In this work, we investigate a weighted hybrid approach to model users' attention to messaging. Through dynamic performance-based weighting, we combine the predictions of three types of models, a general model, a group model and a personalized model to create an approach which can work through the lack of initial data while adapting to the user's behavior. We present the details of our modeling approach and the evaluation of the model with over three weeks of data from 274 users. Our results highlight the value of hybrid weighted modeling to predict when a user cannot attend to their messages.
Data analytics software applications have become an integral part of the decision-making process of analysts. Users of such a software face challenges due to insufficient product and domain knowledge, and find themselves in need of help. To alleviate this, we propose a task-aware command recommendation system, to guide the user on what commands could be executed next. We rely on topic modeling techniques to incorporate information about user's task into our models. We also present a help prediction model to detect if a user is in need of help, in which case the system proactively provides the aforementioned command recommendations. We leverage the log data of a web-based analytics software to quantify the superior performance of our neural models, in comparison to competitive baselines.
Most recommender systems generate recommendations to match the user's current preference. However, users sometimes might have the goal to develop new preferences away from their current preference and use the recommender to guide them towards it. In this paper, we asked users to select a new genre to explore and studied what kind of recommendations would be more helpful for users to start exploring this new music taste. Three different recommendation methods are tested: one non-personalized which recommends the most representative tracks of the genre, one personalized method which considers songs from the new genre that best matches users' current preferences, and one mixed method which makes a trade-off between the two approaches. A comparative design was used in a user experiment in which participants were asked to evaluate the differences between the personalized method/mixed method and the non-personalized baseline. The mixed method results in recommendations that are more accurate and representative for the new genre than the personalized method. Users' perceived helpfulness for exploring the new genre is positively related to both perceived accuracy and perceived representativeness of the recommended items. Besides, recommendations from the mixed method are perceived more helpful for users high on Musical Sophistication Index for Active Engagement (MSAE). To our knowledge, this is one of the first studies using a recommender system to support users' preference development, and provides insights in how recommender systems can help users attain new goals and tastes.
Beyond Explicit Reports: Comparing Data-Driven Approaches to Studying Underlying Dimensions of Music Preference
Prior research from the field of music psychology has suggested that there are factors common to music preference beyond individual genres. Specifically, research has shown that self-reported ratings of preference for individual musical genres can be reduced to 4 or 5 dimensions, which in turn have been shown to correlate to relevant psychological constructs, such as personality. However, the number of dimensions emerging from multiple studies has varied despite the care taken in conducting such research. Data-driven approaches offer opportunities to further this line of research with actual listening data, at a scale and scope surpassing that of traditional psychological studies. Although listening data can be considered more direct and comprehensive evidence of listening preference, transforming this data into meaningful measurements is non-trivial. In the current paper, we report on investigations seeking to find interpretable underlying dimensions of music taste, using implicit large-scale listening data. Offering a critical reflection on potential researchers' degrees of freedom, we adopt an explicit systematic approach, investigating the impact of varying different parameters, analysis, and normalization techniques. More precisely, we consider various ways to extract listening preference information from two large, openly available datasets of music listening behavior, making use of principal component analysis and variational autoencoders to extract potential underlying dimensions. Results and implications are discussed in light of prior psychological theory, and the potential of user listening data to further research on music preference.
Music preferences are likely to depend on contextual characteristics such as location and activity. However, most recommender systems do not allow users to adapt recommendations to their current context. We therefore built ContextPlay, a context-aware music recommender that enables user control for both contextual characteristics and music preferences. By conducting a mixed-design study (N=114) with four typical scenarios of music listening, we investigate the effect of controlling contextual characteristics in a music recommender system on four aspects: perceived quality, diversity, effectiveness, and cognitive load. Compared to our baseline which only allows to specify music preferences, having additional control for context leads to higher perceived quality and does not increase cognitive load. We also find that the contexts of mood, weather, and location tend to influence user perception of the system. Moreover, we found that users are more likely to modify contexts and their profile during relaxing activities.
In this paper, we explore the potential of using visual features in movie Recommender Systems. This type of content features can be extracted automatically without any human involvement and have been shown to be very effective in representing the visual content of movies. We have performed the following experiments, using a large dataset of movie trailers: (i) Experiment A: an exploratory analysis as an initial investigation on the data, and (ii) Experiment B: building a movie recommender based on the visual features and evaluating the performance. The observed results have shown promising potential of visual features in representing the movies and the excellency of recommendation based on these features.
Impact of English Reading Comprehension Abilities on Processing Magazine Style Narrative Visualizations and Implications for Personalization
In this paper, we present research to uncover how the level of reading comprehension abilities impacts how users process textual documents in English with embedded visualizations (i.e., Magazine Style Narrative Visualizations or MSNVs). We analyze performance and gaze data of users processing MSNVs from two user studies, one run in Canada and one in a non-English speaking European country. Our findings provide important insights toward developing automatic, real-time support to MSNV processing personalized according to users' English reading comprehension abilities.
This paper investigates adaptation of feedback to learners' cultural backgrounds. First, we investigate how to portray the cultural background of a learner. Second, we present a qualitative focus-group study, investigating how participants from different cultures believe culture affects the kind of feedback given to a learner. Finally, we present an empirical study on how humans adapt feedback based on the cultural background of learners to inspire an algorithm. Our investigations resulted in a set of stories which can be used to reliably portray a person's culture when investigating cultural adaptation in indirect experiments and user as wizard studies. They also provided insights into the adaptations people make to cultural differences.
Previous research has shown that students from underrepresented minority groups tend to receive lower grades in online classes than their peers, especially in science-focused courses. We propose that there may also be benefits to online courses for these students (e.g., opportunities for peer discussions where minority status is less salient), though little is currently known about these potential benefits. We present a new perspective on learning outcomes by measuring improvement, rather than grades alone. In learning management system data from seven semesters of an online introductory science course, we found that students from underrepresented minority racial groups were indeed less likely to receive high grades, and scored lower on exams; however, their exam scores improved throughout the semester a similar amount compared to their peers. We also compared improvement to students' behaviors, including exam submission times and forum usage, finding that these behaviors were related to improvement. Finally, we also briefly discuss implications of these findings for reducing inequalities in education, and the possibilities for underrepresented minority students in online STEM education in particular.
The success of persuasive systems in changing people's attitudes and behaviours has been established in various domains. Specifically, research has shown that personalized persuasive technology is more effective at achieving the desired goal than the one-size-fits-all approach. However, in the education domain, there are limited studies on the personalization of persuasive strategies to students. To advance persuasive technology research in this area, we investigated the susceptibility of undergraduate students (n = 243) to four persuasive strategies (Reward, Competition, Social Comparison and Social Learning) in order to provide a guideline for designing and personalizing persuasive systems in education. These four strategies were chosen because research on persuasion has established their effectiveness in changing behaviour and/or attitude. The results of our analysis reveal that students are more susceptible to Reward, followed by Competition and Social Comparison (both of which come in the second place) and Social Learning (the least persuasive). Moreover, there is no gender difference in the persuasiveness of the strategies. Therefore, in choosing persuasive strategies to motivate student's learning and success, among the strategies we investigated, Reward should be given priority, followed by Competition and Social Comparison, while Social Learning should be least favoured.
SESSION: Doctoral Consortium
The personalization of feedback by an Intelligent Tutoring System has the potential to greatly improve learner motivation. This PhD investigates how an Intelligent Tutoring System can adapt to the cultural background of learners when giving feedback. The research uses the user-as-wizard method for investigation. To convey the cultural background of the learner in user studies, validated cultural stories (using Hofstede cultural dimensions) are required. These stories are then used to conduct qualitative and empirical studies to investigate how participants from a range of different cultures believe the culture of a learner should affect the kind of feedback given. The insights gathered from these studies will be unified to inspire an algorithm to allow an intelligent tutoring system to utilise these adaptations, and the effects tested on real learners.
Designing Culturally-appropriate Persuasive Technology to Promote Positive Work Attitudes among Workers in Public Workplaces
This research aims to design a mobile persuasive technology (PT) to promote acceptable pro-workplace behaviors and etiquette. As a first step to achieving this, we conducted a user study of 252 subjects from an African organization, to uncover what strategies could be used to model proper behaviors and promote employee's commitment to the ideals, visions and missions of an organization. Leveraging existing workplace behavioral procedures, and socio-cultural strategies, we mapped our findings to their corresponding persuasive techniques. Presently, we employed the iterative design process in developing the mobile PT and the design is informed by our findings. Finally, we will deploy our mobile PT and conduct a large-scale evaluation of public workers in a Nigerian workplace to determine its efficacy to promoting positive workplace etiquette and attitudes. We will employ a mixed-method approach involving both quantitative and qualitative (interview and focus group) for this study.
Since the advent of Web 2.0, Online Social Networks (OSNs) represent a rich opportunity for researchers to collect real user data and to explore OSNs user behaviour. Based on the current challenges and future directions proposed in literature, we aim to investigate how to comprehensively model OSNs user behaviours, by exploiting and combining user data of different nature. We propose to use hypergraphs as a model to easily analyse and combine structural, semantic, and activity-related user information, and to study their evolution over time. This novel user behaviour modelling technique will converge in open, efficient, and scalable libraries, which will be integrated into a modular framework able to handle the data crawling process from several OSNs.
Exploring the Potential of the Resolving Sets Model for Introducing Serendipity to Recommender Systems
Recommender systems offer recommendations based on user's previous ratings. However, sometimes the user is interested in unusual and interesting items that do not exactly match her user profile, as defined by the system. Serendipity, a concept that can be interpreted primarily as surprise, is one of the "beyond-accuracy" aspects that have been proposed to be considered to meet user's expectations for the recommendations she/he gets. Although recent studies attempt to address the serendipity problem, there is still a variety of interpretations regarding the definition, the measurement and the application of serendipity in recommender systems. Our proposed method follows the distance-based approach for multi-dimensional serendipity measurement, which refers to the expected items for the user as a benchmark for measuring serendipity. For integrating serendipity into recommendations, we propose a novel serendipity-oriented user modeling method, based on graph-theory approach - resolving sets in a graph, which enables finding serendipitous items in a multi-dimensional content-based space by detecting the expected items for the user.
UMAP'19 Adjunct: Adjunct Publication of the 27th Conference on User Modeling, Adaptation and PersonalizationFull Citation in the ACM Digital Library
SESSION: Theory , Opinion and Reflection
ACM UMAP - User Modelling, Adaptation and Personalization is the premier international conference for researchers and practitioners working on systems that adapt to individual users, to groups of users, and that collect, represent, and model user information. The Theory, Opinion and Reflection (TOR) track at UMAP is designed to highlight emerging areas of inquiry in UMAP and to promote discussion of potentially visionary ideas.
The spread of radical opinions, facilitated by homophilic Internet communities (echo chambers), has become a threat to the stability of societies around the globe. The concept of choice architecture--the design of choice information for consumers with the goal of facilitating societally beneficial decisions--provides a promising (although not uncontroversial) general concept to address this problem. The choice architecture approach is reflected in recent proposals advocating for recommender systems that consider the societal impact of their recommendations and not only strive to optimize revenue streams. However, the precise nature of the goal state such systems should work towards remains an open question. In this paper, we suggest that this goal state can be defined by considering target opinion spread in a society on different topics of interest as a multivariate normal distribution; i.e., while there is a diversity of opinions, most people have similar opinions on most topics. We explain why this approach is promising, and list a set of cross-disciplinary research challenges that need to be solved to advance the idea.
Human space exploration creates unique challenges and opportunities for many scientific disciplines. From the human-agent interaction perspective, these require significant advances in the way that agents model, adapt and personalize their behavior to individual astronauts and groups of astronauts. In this paper, we highlight the key challenges and opportunities that human space exploration provides to the agent and UMAP communities and present two avenues for future research. We further propose a viable way to explore these challenges and opportunities through the world-wide analogue space programs which solicit research proposals from all scientific disciplines.
Human cognitive biases are numerous and well established. Due to inherent limitations in our knowledge of the world, and computational constraints, our judgments and decisions do not rigidly adhere to the principle of maximizing expected utility. We frequently employ cognitive shortcuts, ignoring relevant information, and make errors in how we store and retrieve items from memory. Human decisions are additionally influenced by moral, emotional and cultural parameters. People often perceive value in a way that is very different from well-established decision-theoretic frameworks, but much of the work on personalization does not capture human cognitive biases. Our central hypothesis is that a new generation of recommendation systems can be designed by explicitly modeling human cognitive biases such as contrast, decoy, distinction, and framing. We are just now beginning to see explicit non-linear models of human risk perception being incorporated into machine learning algorithms, and we believe this trend will accelerate in the near future. In this paper we review today's recommendation systems, give an analysis of their limitations and make an argument for why future recommendation systems should incorporate explicit models of human cognitive bias.
SESSION: Demo and Late-breaking Results
It is our great pleasure to welcome you to the UMAP 2019 LBR and Demo Track, in conjunction with the 27th Conference on User Modelling, Adaptation and Personalization, held in Larnaca, Cyprus on June 9-12th, 2019.
This track encompasses two categories: (i) Demos, which showcase research prototypes and commercially available products of UMAP-based systems, (ii) Late-breaking Results (LBR), which contain original and unpublished accounts of innovative research ideas, preliminary results, industry showcases, and system prototypes, addressing both the theory and practice of UMAP.
The submissions spanned a wide scope of topics, ranging from novel techniques for user and group modeling, to adaptation and personalization implementations across different application scenarios.
We received 46 LBR and 4 Demo submissions. Each submission was carefully reviewed by members of the Demo and LBR program committee, which consisted of 89 members. Each submission was reviewed by at least 3 PC members.
Out of this total of 50 submissions, 15 LBR and 3 Demos were deemed of good quality by the reviewers, and were consequently accepted (36% overall acceptance rate). They were presented in the UMAP poster sessions, which collectively showcased the wide spectrum of novel ideas and latest results in user modelling, adaptation and personalization.
This paper introduces 'RehaBot', a framework for building adaptive serious games in the context of telerehabilitation. RehaBot takes advantage of 3D motion tracking and virtual reality devices, to develop an immersive and gamified telerehabilitation environment. A personalized and adaptive gaming system is developed, which allows patients to perform exercises with the help of embedded virtual assistants, hereafter called 'rehab bots', that are dynamically displayed within scenes to guide the patient through the different sets of gestures required to complete the session. These rehab bots have the ability to learn and adapt to the best level of difficulty in real-time based on the user performance. An intelligent alerting and automatic correction technique is incorporated within our engine, so that pre-calculated gesture patterns are correlated and matched with patients' gestures. Consequently, the system estimates the perceived difficulty of gestures by the patient, and automatically adjusts the game behavior to ensure a highly engaging and adaptive gaming experience. Furthermore, multimodal instructions are conveyed to users with details on joints that are not performing as expected, and to guide them towards improving the current gesture. A pilot study has been conducted to prove the usability and effectiveness of our adaptive physiotherapy solution.
In this paper we present a semantics-aware recommendation strategy that uses graph embedding techniques to learn a vector space reresentation of the items to be recommended. Such a representation relies on the tripartite graph which connects users, items and entities gathered from DBpedia, thus it encodes both collaborative and content-based information. These embeddings are then used to feed with positive and negative examples (the items the user liked and those she did not like) a classification model, which is finally exploited to classify new items as interesting or not interesting for the target user. In the experimental evaluation we evaluate the effectiveness of our method on varying of different graph embedding techniques and on several topologies of the graph. Results show that the embeddings learnt by combining collaborative data points with the information gathered from DBpedia led to the best results and also beat several state-of-the-art techniques.
Due to the rise of available online music, a lot of music consumption is moving from traditional offline media to online sources. Online music sources offer almost an unlimited music collection to its users. Hence, how music is consumed by users (e.g., experts) may differ from traditional offline sources. In this work we explored how musically sophisticated users (i.e. experts) consume online music in terms of diversity. To analyze this, we gathered data from two different sources: Last.fm and Spotify. As expertise is defined by the ubiquitousness of experiences, we calculated different diversity measurements to explore how ubiquitous (in terms of diversity) the listening behaviors of users are. We found that different musical sophistication levels correspond to applying diversity related to specific kind of musical characteristics (i.e., artist or genre). Our results can provide knowledge on how systems should be designed to provide better support to expert users.
Onboarding users to a complex application or a new functionality can be a serious issue, especially for organizations that need to train their new employees. Using a complex application without proper training or guidance can lead to users' confusion and frustration. In this paper, we introduce the onboarding platform YesElf intended for web applications. Its approach to onboarding is to use embedded guides within the application; its novelty lies in the robustness, ease of setup and integration of the YesElf guides into any web-based application. Most importantly, YesElf supports personalized adaptation of user guidance. This, we demonstrate by a novel method for automated recognition of user's confusion in real time that we integrated into YesElf. The information on user's confusion serves as a basis for adaptive display of the guides, when they are needed the most. We evaluated the proposed method on the data collected in a user study with 60 participants and achieved 63% precision which outperforms the state-of-the-art classifier based on the eye tracking data (although, in our case, we used the more readily available mouse movement data).
This paper studies an approach for predicting an individual's perception of the advertising appeal of ad design. Although previous research has shown that people perceive an ad as more appealing when its design matches their psychological traits, the matching required the help of psychology experts, relying on their implicit knowledge. To exclude such dependence, we examined how the psychological traits affect perceived advertising appeal by conducting a questionnaire survey. Analyzing the survey results, we confirmed that psychological traits have significant moderating effects on both visual and linguistic features of an ad design in terms of how an individual perceives advertising appeal. We also confirmed that the moderating effects as well as main effects of visual and linguistic features have significant predictive utility for perceived appeal. The model in which the both effects are incorporated predicted the most appealing ad for each person more accurately than did a human without psychology expertise (accuracy lift: mean - 1.35, max - 2.48). While further study is necessary on whether the studied approach can serve as a substitute for psychology experts, we consider the present study took the first step toward realizing this goal.
The world is changing and with the evolution of technology computers have become an essential part of humans' lives. Nowadays many people use computers for both work and personal life, and especially spending hours sitting at a desk in front of their computer screens. This phenomenon negatively influences people's health, affecting their skeletal and ocular systems. As a result, several different ergonomic solutions have been suggested to address these challenges. This paper proposes a solution which adjusts the computer screen position, elevation and orientation in order to reduce the physical load on the user and better fit it to their posture.
Socially Responsive eCommerce Platforms: Design Implications for Online Marketplaces in Developing African Nation
Before the advent of internet technologies, African merchants have always had indigenous none techno-driven strategies with which they use to engage, maintain and satisfy customers within their communities. However, due to the cross-border marketing/economic opportunities that the internet provides, many African merchants are setting up eCommerce sites for their businesses. Research shows that successful eCommerce sites are operationalized with various persuasive techniques to promote customer engagement and improve user experiences. This is also true for African based eCommerce sites. Various researchers have evaluated and compared persuasive techniques operationalized on indigenous African eCommerce sites. However, none of those studies sought to understand the implications of the age-long traditional marketing strategies in the design of African based eCommerce platforms. Therefore, this paper analyzes the persuasive techniques employed in conventional African marketplaces to uncover the design requirements that could be operationalized on the eCommerce version. To achieve this objective, we conducted a mixed method study on 151 participants. We conducted qualitative studies (comprising interviews, observations, and conceptual investigations) on 50 African merchants to uncover the techniques they use to attract, satisfy and retain their customers in the conventional market. Secondly, we conducted quantitative studies on 101 customers to uncover the effects of those techniques on them. Among other things, the results from the studies revealed various techniques, which traditional African merchants use to attract and retain customers. The results also revealed the effects of those techniques on customers' purchasing behaviors. We offer design guidelines to operationalize and tailor those techniques to attract, recruit, satisfy, retain and promote repeated purchases on eCommerce that target customers from developing countries.
User profiling is becoming increasingly holistic by including aspects of the user that until a few years ago seemed irrelevant. The content that users produce on the Internet and social networks is an essential source of information about their habits, preferences, and behaviors in many situations. One factor that has proved to be very important for obtaining a complete user profile that includes her psychological traits are the emotions experienced. Therefore, it is of great interest to the research community to develop approaches for identifying emotions from the text that are accurate and robust in situations of everyday writing. In this work, we propose a classification approach based on deep neural networks, Bi-LSTM, CNN, and self-attention demonstrating its effectiveness on different datasets. Moreover, we compare three pre-trained word-embeddings for words encoding. The encouraging results obtained on state-of-the-art datasets allow us to confirm the validity of the model and to discuss what are the best word embeddings to adopt for the task of emotion detection. As a consequence of the great importance of deep learning in the research community, we promote our model as a starting point for further investigations in the domain.
In many settings, it is required that items are recommended to a group of users instead of a single user. Most often, when the decision criteria and preferences of the group as a whole are not known, the gold standard is to aggregate individual member preferences or recommendations. Such techniques typically presuppose some process under which group members reach consensus, e.g., least misery, maximum satisfaction, disregarding any uncertainty on whether this presumption is accurate. We propose a different approach that explicitly models the system's uncertainty in the way members might agree on a group ranking. The basic idea is to quantify the likelihood of hypothetical group rankings based on the observed member's individual rankings. Then, the systems recommends a ranking that has the highest expected reward with respect to the hypothetical rankings. Experiments with real and synthetic groups demonstrate the superiority of this approach compared to previous work based on aggregation strategies and to recent fairness-aware techniques.
We analyze the effect of a smile in personas pictures on persona perceptions, including credibility, likability, similarity, and willingness to use. We conduct an online experiment with 2,400 participants using a 16-item survey and multiple persona profile treatments of which half have a smiling photo and half do not. We find that persona profiles with a smiling photo result in an increase in perceived similarity with, likability of, and willingness to use the personas. In contrast, a smile does not increase the credibility of the personas. Our research has implications for the design of persona profiles and adds to previous findings of persona research that the picture choice influences individuals' persona perceptions in profound ways.
Integrated Development Environments (IDEs) are used for a varietyof software development tasks. Their complexity makes them chal-lenging to use though, especially for less experienced developers. In this paper, we outline our approach for an user-adaptive IDE that is able to track the interactions, recognize the user's intent and expertise, and provide relevant, personalized recommendations in real-time. To obtain a user model and provide recommendations, interaction data is processed in a two-stage process: first, we derive a bandit based global model of general task patterns from a dataset of labeled interactions. Second, when the user is working with the IDE, we apply a pre-trained classifier in real-time to get task labels from the user's interactions. With those and user feedback we fine-tune a local copy of the global model. As a result, we obtain a personalized user model which provides user-specific recommendations. We finally present various approaches for using these recommendations to adapt the IDE's interface. Modifications range from visual highlighting to task automation, including explanatory feedback.
Using Personal Assistants (PAs) via voice becomes increasingly usual as more and more devices in different environments offer this capability, such as Google Assistant, Amazon Alexa, Apple Siri, Microsoft Cortana, Mercedes-Benz MBUX or BMW Intelligent Personal Assistant. PAs help users to set reminders, find their way through traffic, or send messages to friends and colleagues. While serving the users' needs, PAs constantly collect personal data in order to personalize their services and adapt their behavior. In order to find out which objective Cognitive Load (CL) indicators reflect the users' perception of proactive system behavior in six specific use cases of an in-car PA, we conducted a Wizard of Oz study in a driving simulator with 42 participants. We varied traffic density and tracked physiological data, such as heart rate (HR) and electrodermal activity (EDA). We assessed the users' CL during the interaction with the PA by employing these data as well as real-time driving data (RTDA) via the Controller Area Network (CAN bus). The results show that physiological data like HR and EDA are not suitable to be used as indicators for the users' CL in this experiment. This is because the tracked physiological data do not show significant differences with respect to different traffic densities or proactivity. At the same time it has to be discussed whether the used type of recording physiological data is robust enough for our purposes. Concerning driving data, only the acceleration parameter showed a tendency towards differences between age groups, though insignificantly. The same is valid for the steering angle parameter when comparing male and female users. For future work, we plan to additionally evaluate subjective CL measures and other ratings to see whether these show more significant differences between the (non-)proactive assistants, traffic densities, or use cases.
In group recommender systems,decision manipulation refers to an attack in which a group member makes attempts to push his/her favorite options. In this paper, we propose user interfaces to counteract decision manipulation in group recommender systems. The proposed user interfaces visualize information dimensions regarding rating adaptations of group members at different transparency levels. The results show that the user interface at the highest transparency level best helps to discourage users from decision manipulation. Besides, the ability of the user interfaces to counteract decision manipulation differs depending on the dimensions represented in the user interfaces. The information dimensions regarding "\textititem ratings " and "\textitgroup recommendations " have the strongest impacts on preventing users from decision manipulation.
Exer-model: A User Model for Scrutinising Long-term Models of Physical Activity from Multiple Sensors
A user model that is built from the data of multiple physical activity sensors has the potential to enable people to answer important questions about their long-term physical activity. Our work provides a way to do this for the case of exercise from virtual reality gaming and from incidental daily walking. Our approach is based two parts: 1) a carefully designed a user model ontology, Exer-model; 2) an interface for navigating the model and comparing components of the model. We evaluated the Exer-model ontology and the scrutiny interface in a study with 16 users: 8 viewing their own user models, from 8 weeks of their sensor data, and the other 8 scrutinising the model of a hypothetical user. Our core contributions are the insights about designing the ontologies and interfaces for scrutable user models from multiple physical activity sensors.
While software tools can be very powerful, a one-size-fits-all approach does not work because individual needs vary over time. Co-Adapt is a software framework that learns from user activity and adjusts the interface in real time to suit changing needs. Prior research has shown that the best in class user interfaces (UI) are not as effective across multiple groups of people. We performed three experiments by integrating with three distinct UI prototypes that tested iteratively tailored software to show the effectiveness of different UI presentations. Co-Adapt demonstrates that a multitude of UIs improve usability, which in some instances lead to greater tool adoption. Our success across three domains suggests generalizability of the framework and is promising for further experimentation across other application areas.
Zero-Coding UMAP in Marketing: A Scalable Platform for Profiling and Predicting Customer Behavior by Just Clicking on the Screen
Customer Data Platform (CDP) is an integrated customer database operated by marketers. In the context of UMAP, this paper demonstrates a real-world CDP with a special focus on (1) simple and deterministic text-based behavioral profiling technique, and (2) GUI-based versatile tool for predictive analytics. Those functionalities are designed for those who have no expertise in machine learning and natural language processing, so the only thing marketers have to do is clicking some buttons on UI. Meanwhile, their back-end system ensures scalability and utility of the entire workflow from data collection and management to prediction and visualization.
This paper briefly describes aspects of the Tikkoun Sofrim crowdsourcing webApp. Tikkoun Sofrim is a webApp which allows users to correct automatic transcriptions (AT) done by an AI Neural network engine. We look at the background of the crowdsourcing phenomenon in the use of automatic transcription of digital humanities documents. System structure is briefly described. We then examine personalization and adaption aspects at different stages of the user/application lifecycle Finally, we briefly list future challenges.
SESSION: Workshops and Tutorials
It is our great pleasure to welcome you to the ACM UMAP 2019 Workshops and Tutorials proceedings. In the 27th edition of the ACM Conference on User Modeling, Adaptation and Personalization, a total of 10 workshop and 3 tutorial proposal submissions were received for consideration. After a single-blind peer review process, 7 workshops and 2 tutorials were accepted. They were held in conjunction with the conference in Larnaca (Cyprus) on the 9th of June, 2019.
SESSION: Adaptive and Personalized Persuasive Technology (ADAPPT 2019) Workshop
UMAP 2019 ADAPPT (Adaptive and Personalized Persuasive Technology) Workshop Chairs' Welcome & Organization
We are pleased to welcome you to the 1st International Workshop on Adaptive and Personalized Persuasive Technology (ADAPPT 2019). ADAPPT 2019 is a half-day workshop held in conjunction with the 27th ACM Conference on User Modeling, Adaptation and Personalization (UMAP 2018), 09-12 June 2019 in Larnaca, Spain. Persuasive technologies are increasingly being used to bring about behavior change in various domains of human endeavors, including health, education, commerce, energy conservation, safety, etc. However, research on personalizing and adapting them to their target users to make them more effective is still in its infancy. As such, for the first time, we proposed at the ACM UMAP 2019 conference the ADAPPT 2019 workshop. The workshop aims to bring together researchers and practitioners from academia and industry-working in the area of adapting and personalizing persuasive technologies-to present, discuss and share their work in progress with other members of the research community. Specifically, the workshop aims to provide a platform for stakeholders to brainstorm, identify and discuss the opportunities and challenges in the ADAPPT field as well as emerging techniques, methods and approaches to personalizing and adapting persuasive technologies to the target users.
In the first edition of the workshop, we received 10 submissions from four different countries, including Canada, Germany, Nigeria and Spain, covering a wide range of topics in domains such as health, education, organization, social media, e-commerce, etc. Each of the 10 papers was reviewed by at least two reviewers, which included members of the organizing committee and external reviewers with expertise in different areas of persuasive technology research. All of the 10 papers, which include 4 full papers and 6 short papers, were accepted for presentation at the workshop.
Social Network Sites (SNSs) like Facebook or Instagram are spaces where people expose their lives to wide and diverse audiences. This practice can lead to unwanted incidents such as reputation damage, job loss or harassment when pieces of private information reach unintended recipients. As a consequence, users often regret to have posted private information in these platforms and proceed to delete such content after having a negative experience. Risk awareness is a strategy that can be used to persuade users towards safer privacy decisions. However, many risk awareness technologies for SNSs assume that information about risks is retrieved and measured by an expert in the field. Consequently, risk estimation is an activity that is often passed over despite its importance. In this work we introduce an approach that employs deleted posts as risk information vehicles to measure the frequency and consequence level of self-disclosure patterns in SNSs. In this method, consequence is reported by the users through an ordinal scale and used later on to compute a risk criticality index. We thereupon show how this index can serve in the design of adaptive privacy nudges for SNSs.
Competition has been identified as an intrinsic motivation that could lead to successful outcomes in education. However, in Persuasive Technology in Education (PTE) research, there are limited studies showing its possible predictors. To advance research in this area, we conducted an empirical study among university students (N = 243) to uncover how extrinsic factors and social influence, which are external to learners, influence students' susceptibility to Competition. Specifically, we investigated how Social Learning, Social Comparison, and Reward, which are widely applied in persuasive technologies (PTs), influence Competition. Our results show that Social Comparison and Reward have significant influence on Competition, with Social Comparison (β = 0.52, p < 0.001) having a stronger influence than Reward (β = 0.29, p < 0.001). However, Social Learning (β = -0.07, p = n.s) has no significant effect on Competition. Our model accounts for about 41% of the variance of Competition. Moreover, our multigroup analysis reveals that there are no significant differences between males and females, indicating that our findings generalize across gender. These findings suggest that Social Comparison, Reward, and Competition are compatible strategies, which can be implemented together in a persuasive system for education. We discuss the implications of our findings
Susceptibility to Fitness App's Persuasive Features: Differences Between Acting and Non-Acting Users
The incidence of physical inactivity, obesity and non-communicable diseases is on the rise globally due to the sedentary lifestyles occasioned by modernity and technology. As a means of tackling the inactivity problem, which is almost becoming a global epidemic, research has shown that persuasive technology holds bright prospects. However, in the physical activity domain, there is limited research on users' persuasion profiles and the differences between users who are currently exercising (acting users) and those who have the intentions to exercise in the future (non-acting users). To bridge this gap, we conducted a study among 190 participants resident in two individualist countries to determine the susceptibility profile of both user types and their differences. We based our study on storyboards, illustrating six commonly employed persuasive features in fitness apps. The results of our analysis showed that both user types are most likely to be susceptible to Goal-Setting/Self-Monitoring, followed by Reward and Competition, and least likely to be susceptible to Cooperation, Social Comparison and Social Learning. In particular, acting users are more likely to be susceptible to Social Learning than non-acting us-ers. Overall, our findings suggest that, irrespective of user type, personal features will be more likely effective than social features among users from individualist cultures. We discuss the implications of our findings in the context of fitness apps design.
Understanding a consumer's motivation to shop online with a vendor can help an e-business better understand the attitude of customers and what they look out for in their shopping decision-making process. Equally important in the shopping decision making process is the influence of the perceived quality of products and their price. Understanding how consumers are influenced by the perceived quality and price of products can help e-businesses to improve their customers' shopping experience. To contribute to ongoing research in this area, we investigate the influence of perceived product quality and price on the motivation of e-shoppers to shop online. In particular, we investigate which of perceived quality and price have a greater influence on the consumer's motivation to shop online. We also investigate the moderating effect of income and gender. Using a sample size of 241 e-commerce shoppers, we develop and test a global research model using Partial Least Squares-Structural Equation Modeling (PLS-SEM). Our results suggest that balanced buyers (shoppers who are moderately motivated by convenience and variety seeking but do not plan ahead and are impulse buyers) are more influenced by the relative price of products compared to their quality. In addition, balanced buyers who earn over $30,000 are influenced by the quality of the product compared to those who earn less than $30,000. Furthermore, male shoppers who are motivated by the convenience of online shopping (convenience shoppers) are also influenced by the perceived quality of products compared to female shoppers who are not.
Online controlled experiments (also called A/B tests, bucket testing or randomized experiments) have become an habitual practice in numerous companies for measuring the impact of new features and changes deployed to softwares products. In theory, these experiments are one of the simplest methods to evaluate the potential effects that new features have on user's behavior. In practice, however, there are many pitfalls that can obscure the interpretation of results or induce invalid conclusions. There is, in the literature, no shortage of prior work on online controlled experiments addressing these pitfalls and conclusions misinterpretations, but the topic is not tackled considering the specific case of testing personalization features. In this paper, we present some of the experimentation pitfalls that are particularly important for personalization features. To better illustrate each pitfall, we include a combination of theoretical argumentation as well as examples from real company's experiments. While there is clearly value in evaluating personalized features by means of online controlled experiments, there are some pitfalls to bear in mind while testing. With this paper, we aim to increase the experimenters' awareness of leading to improved quality and reliability of the results.
Physical inactivity has been recognized as one of the leading causes of non-communicable diseases and mortality globally. Though persuasive technology has been identified as a potential tool for tackling physical inactivity and sedentary behaviors, very little attention has been paid to investigating the effectiveness of culture-tailored interventions in the wild. To bridge this gap, we designed and implemented two versions of a fitness app we called BEN'FIT [personal version (PV) and social version (SV)] targeted at encouraging regular bodyweight exercise behavior on the home front. The PV and SV versions are targeted at users from individualist and collectivist cultures, respectively. In this paper, we describe the empirical findings that informed the design and implementation of both versions of the BEN'FIT app, their features and how we intend to evaluate them in a pilot field study among our target audience once we complete the implementation of the app.
How Effective Are Social Influence Strategies in Persuasive Apps for Promoting Physical Activity?: A Systematic Review
The use of behavior change systems and persuasive technologies to promote desirable behavior is increasingly gaining attention. Most existing Persuasive Technologies (PTs) are targeted at promoting Physical Activity (PA) using three common socially-oriented persuasive strategies: competition, social comparison, and cooperation. This paper provides an empirical review of 19 years (54 papers) of literature on persuasive technology for physical activity promotion. The review aims to (1.) evaluate the effectiveness of PTs employing social influence strategies to promote PA; (2.) summarize and highlight trends in the outcomes and employed technological platforms; (3.) reveal some weaknesses of existing PTs for promoting PA; and finally, (4.) offer suggestions for improvements, and opportunities for future research in this area.
Recovering from Stroke can be a very long and stressful process. It may involve several months or years of exercise and other medical routines that could be both painful and uninteresting. Most people suffering from stroke tend to lose their basic motor functions and it may take a series of exercises to gain back their full manual dexterity. There is a need for a variety of interventions to make these exercises engaging and exciting, to ease the burden of stroke rehabilitation. This paper explores the possibility of using portable personal electroencephalogram (EEG) devices with persuasive games as a tool for stroke rehabilitation. It looks at the major limitations and strengths of using personal Brain-Computer Interfaces (BCIs) for stroke rehabilitation. The paper also presents the design and development of a Brain-Computer Interface persuasive game called Rock Evaders, that aims to motivate people recovering from stroke to carry out their rehabilitation exercises in an exciting and engaging way.
Developing Persuasive Mobile Games for African Rural Audiences: Challenges implementing the Persuasive Techniques
With the advent of cheap Android phones in the African Tech market, most people living in the rural areas of many African countries now have access to smartphones. These phones give them the opportunity to have an almost identical mobile phone experience as people living in urban areas or the Western world. This development opens a window of opportunity to leverage this high penetration of mobile devices to design application such as persuasive game interventions to assist individuals living in these communities to modify, change or shape their behaviours and attitudes in a desirable way. This paper explores the challenges and issues encountered in the design and use of persuasive mobile games as a tool to promote behaviour change among people living in the Rural African communities. It also highlights how these challenges affect the implementation of persuasive strategies, suggests design solutions for overcoming these challenges, and how persuasive games can be optimized to be appropriate for the target rural African populations. Some of these challenges are technically oriented (internet connectivity issues) while others are non-technically oriented (language diversity).
This paper investigates how mobile persuasive system targeting African audience could be designed and tailored to promote employee's commitment to the ideals, visions, and mission of an organization. We conduct a qualitative study with two categories of workers to uncover core factors that influence employee's attitudes to their jobs and map our findings to their matching social influence persuasive techniques. We propose that a persuasive system (PS) employing the social influence strategies could motivate workers towards acceptable positive pro-workplace behaviors and etiquette. The PS allows workers too compare their behaviors against set goals and acceptable standards, compete and compare performances with peers, view and respond to peers' activities, and receive recognition for accomplishing a target task. The system ensures the security of worker's data via the authentication of login credentials while showing them a personalized persuasive display of essential workplace information. We present a prototype persuasive system called "PAULApp" for motivating pro-workplace behaviors and plans for evaluation. PAULApp was designed using the iterative design process and was informed by the findings from the user studies.
SESSION: Adaptive and Personalized Privacy and Security (APPS 2019) Workshop
Adaptive and Personalized Privacy and Security (APPS 2019): Workshop Chairs' Welcome and Organization
It is our great pleasure to welcome you to the First International Workshop on Adaptive and Personalized Privacy and Security (APPS 2019). APPS 2019 (http://appsworkshop.cs.ucy.ac.cy) is a half-day workshop held on June 09, 2019, in conjunction with the ACM Conference on User Modeling, Adaptation and Personalization (ACM UMAP 2019) in Larnaca, Cyprus. Adaptive and personalized privacy and security aims at supporting privacy- and/or security-related tasks by leveraging on holistic user models which reflect the users' unique sociocultural, physical, physiological and technological context in which interaction takes place. As such, APPS 2019 aims to bring together researchers and practitioners working on diverse topics related to understanding and improving the usability of privacy and systems security, by applying user modeling, adaptation and personalization principles framed by User-Centered Design methods.
This paper studies the availability of apps and app stores across countries. Our research finds that users in specific countries do not have access to popular app stores due to local laws, financial reasons, or because countries are on a sanctions list that prohibit foreign businesses to operate within its jurisdiction. Furthermore, this paper presents a novel methodology for querying the public search engines and APIs of major app stores (Google Play Store, Apple App Store, Tencent MyApp Store) that is cross-verified by network measurements. This allows us to investigate which apps are available in which country. We primarily focused on the availability of VPN apps in Russia and China. Our results show that despite both countries having restrictive VPN laws, there are still many VPN apps available in Russia and only a handful in China. In addition, we have included findings of a global search for the availability of privacy-enhancing and other apps that are known to be censored. Finally, we observe that it is difficult to find out which apps have been removed or are unavailable on the examined app stores. As a consequence, we urge all app store providers to introduce app store transparency reports, which would include when apps were removed and for what reasons.
On the Personalization of Image Content in Graphical Passwords based on Users' Sociocultural Experiences: New Challenges and Opportunities
Recent works underpin the added value of considering users' past sociocultural experiences as a personalization factor for the image content used within graphical password schemes, since it has a positive impact on the security and memorability of the user-chosen passwords. This paper discusses the need for personalization of the image content used in graphical password schemes, as well as the initial steps towards the realization of an image content personalization framework that aims to achieve a better equilibrium between security and memorability. The paper also discusses emerging challenges related to the elicitation and maintenance of individual sociocultural-centered user models, the image content personalization mechanism and privacy considerations.
Passwords are still a widespread authentication mechanism that, despite the efforts of the cybersecurity community to educate people, are often predictable. Therefore, there is a need for defenders, e.g. cybersecurity/IT administrators, to periodically assess the users' passwords in their organization, improve their awareness on the security level and take measures to improve the situation. Password cracking can assist in the evaluation of the strength of passwords and a variety of tools exist to execute it. The challenge with this is that it is a time-consuming process and it needs to be optimized to detect weak passwords within a specific evaluation timeframe. To optimize the process, knowledge in the area and appropriate tools are required. However, even though a lot of research is performed in this area, the knowledge and tools are scarce, challenging defenders' tasks. Therefore, the need arises to promote the design of advanced tools, integrating existing user knowledge and creating powerful toolkits. This work presents the design of UPAT (Ultimate Password Awareness Toolkit), which specifies essential features to optimize the password cracking process. The evaluation results are encouraging as to the tool's effectiveness and users' satisfaction, demonstrating the importance of designing next generation password cracking toolkits to enhance the security of communication and information systems.
Adaptive systems are based on user preferences and needs, while sensing and context awareness are also essential features. Recently, the notion of smart homes requires the collection of different data from users, in order to provide them with a personalized user experience. In this context, smart water management can facilitate activities within the smart home. User privacy protection is vital in this environment to provide adaptive data collection and usage. In this paper, we introduce our vision towards user privacy protection in this setting by specifying and considering user privacy preferences, and we present our ongoing work and its initial results. The current work is being conducted in the framework of the TAMIT research project. This initial work will serve as a basis for the integration of user privacy management within TAMIT and can be a useful source of information for the management of user privacy preferences in similar platforms.
Security And Privacy Of Medical Data: Challenges For Next-Generation Patient-Centric Healthcare Systems
We describe the recently-started EU H2020 Serums: Securing Medical Data in Smart Patient-Centric Healthcare Systems project that aims to develop novel techniques for safe and secure collection, storage, exchange and analysis of medical data, allowing the patients of the next-generation smart healthcare centers to get the best pos- sible treatment while respecting privacy and ownership of their sensitive personal data. Our goal is to signi cantly enhance trust in the new medical systems. We outline the techniques that will be extended/developed over the course of the project and describe the use cases that will be used to verify the e ectiveness of these technologies in practice.
SESSION: Towards Comparative Evaluation in user Modeling, Adaptation and Personlization (EvalUMAP 2019) Workshop
Research in the areas of User Modelling, Adaptation and Personalization faces a number of significant scientific challenges. One of the most significant of these challenges is the issue of comparative evaluation. It has always been difficult to rigorously compare different approaches to personalization, as the function of the resulting systems is, by their nature, heavily influenced by the behavior of the users involved in trialing the systems. Developing comparative evaluations in this space would be a huge advancement as it would enable shared comparison across research. This workshop aims to develop initial shared task(s) that will be published ahead of UMAP 2020, providing opportunity for participants to test and tune their systems and complete the task in order for comparative results and associated publications to be prepared for and presented at UMAP 2020.
Traditional recommender systems present a relatively static list of recommendations to a user where the feedback is typically limited to an accept/reject or a rating model. However, these simple modes of feedback may only provide limited insights as to why a user likes or dislikes an item and what aspects of the item the user has considered. Interactive recommender systems present an opportunity to engage the user in the process by allowing them to interact with the recommendations, provide feedback and impact the results in real-time. Evaluation of the impact of the user interaction typically requires an extensive user study which is time consuming and gives researchers limited opportunities to tune their solutions without having to conduct multiple rounds of user feedback. Additionally, user experience and design aspects can have a significant impact on the user feedback which may result in not necessarily assessing the quality of some of the underlying algorithmic decisions in the overall solution. As a result, we present an evaluation framework which aims to simulate the users interacting with the recommender. We formulate metrics to evaluate the quality of the interactive recommenders which are outputted by the framework once simulation is completed. While simulation alone is not sufficient to evaluate a complete solution, the results can be useful to help researchers tune their solution before moving to the user study stage.
This paper presents an approach and methodology for user-centred evaluation of adaptive systems. In contrast to layered evaluation approaches that decompose adaptation into its constituents, our approach conceptualises the quality and benefit for the user into separate evaluation qualities for a comprehensive and multifaceted evaluation. Instruments of different modalities are used to measure these qualities from the user perspective. A service is presented that takes up this approach and enables time- and cost-efficient evaluation by defining and re-using evaluation qualities and instruments, as well as collecting and analysing data based on these definitions. This approach allows to compare different adaptive systems by using the same qualities and adapting the instruments to the specific characteristics of the particular adaptive system.
The Conlan whitepaper makes a compelling case for a community-scale means of evaluating algorithms for user modeling, adaptation and personalisation (UMAP). The authors propose an evaluation paradigm focused on a personalisation use case within an open modeling environment. Their use case is one where a mobile interface learns a user's preferences for different notifications in different contexts. Additional use cases could be incorporated in this paradigm to provide contrasting kinds of challenge for personalisation evaluation. In this position paper we summarise a use case where personalisation is achieved by categorising a user and switching their interface to a variant. The user interface is a mobile news app within a platform that also comprises a user modeling function and an interface personalisation service. Comparison of the use cases helps to map the space for the evaluation paradigm.
A shared challenge in the domain of User Modeling, Adaptation and Personalisation is proposed for the 2019 EvalUMAP workshop whereby the evaluation of user models generating personalised push-notifications is to be explored. As such, this paper presents a description of the evaluation process, a solution to the first proposed challenge, a discussion of results obtained from the Gym-Push evaluation environment and a number of benchmarks which can be used as a baseline for future work.
SESSION: Explainable and Holistic User Modeling (ExHUM 2019) Workshop
It is our great pleasure to welcome you to the UMAP 2019 Workshop on Explainable and Holisitic User Modeling (ExHUM). Our workshop took inspiration from the analysis of the recent Web dynamics: according to a recent claim by IBM, 90% of the data available today have been created in the last two years. Such an exponential growth of personal information has given new life to research in the area of user modelling, since information about users' preferences, sentiment and opinions, as well as signals describing their physical and psychological state, can now be obtained by mining data gathered from many heterogeneous sources. How can we use such data to drive personalization and adaptation mechanisms? How can we effectively merge such data to obtain a holistic representation of all (or some of) the facets describing people? Moreover, as the importance of such technologies in our everyday lives grows, it is also fundamental that the internal mechanisms that guide personalization algorithms are as clear as possible. It is not by chance that the recent General Data Protection Regulation (GDPR) emphasized the users' right to explanation when people face machine learning-based systems. Unfortunately, the current research tends to go in the opposite direction, since most of the approaches try to maximize the effectiveness of the personalization strategy (e.g., recommendation accuracy) at the expense of the explainability and the transparency of the model.
Other than private broadcasters, publicly financed broadcasters have to fulfil a public service remit. Individual playouts in public radio, therefore, consist not only of recommender content but also of 'anti-recommender content" that matches public interests. Such anti-recommender content in individual playouts may be unexpected for users and may need explanation. To find out what explanations might look like in public radio, we elicit the requirements of the public service remit for an example country. Based on these requirements, we propose an approach for designing explanations of recommendations that align with the public service remit.
Social Tags and Emotions as main Features for the Next Song To Play in Automatic Playlist Continuation
The broad diffusion over the Internet of songs streaming services points out the need for implementing efficient and personalized strategies for incrementing the fidelity of the customers. This scenario can collect enough information about the user and the items for successfully design a Recommender System for the automatic continuation of playlists of digital contents. In particular, in this work we proposed a strategy for suggesting a set of tracks, starting from a list of songs played by the user, candidate as next to play. The list contains songs that are coherent with the main characteristics of songs already played. In order to collect enough information and for applying a recommendation strategy, we used third-party external sources of information. They provide data about the song, including its popularity, the emotion evoked by its lyrics, low and high-level audio features, lyrics and more. The system highlights the importance to use user-generated tags and emotional features for successfully predicts user next played songs.
Personalized content provided by recommender systems is an integral part of the current online news reading experience. However, news recommender systems are criticized for their 'black-box' approach to data collection and processing, and for their lack of explainability and transparency. This paper focuses on explaining user profiles constructed from aggregated reading behavior data, used to provide content-based recommendations. The paper makes a first step toward consolidating epistemic values of news providers and news readers. We present an evaluation of an explanation interface reflecting these values, and find that providing users with different goals for self-actualization (i.e., Broaden Horizons vs. Discover the Unexplored) influences their reading intentions for news recommendations.
Many explainable recommender systems construct explanations of the recommendations these models produce, but it continues to be a di cult problem to explain to a user why an item was recommended by these high-dimensional latent factor models. In this work, We propose a technique that joint interpretations into recommendation training to make accurate predictions while at the same time learning to produce recommendations which have the most explanatory utility to the user. Our evaluation shows that we can jointly learn to make accurate and meaningful explanations with only a small sacri ce in recommendation accuracy. We also develop a new algorithm to measure explanation delity for the interpretation of top-n rankings. We prove that our approach can form the basis of a universal approach to explanation generation in recommender systems.
Attention plays an important role in the daily lives of people and especially in learning. It may be influenced by different factors since childhood or during the aging process because of deficits such as Attention Deficit Hyperactivity Disorder (ADHD), which is responsible of learning difficulties. Serious games for attention training may be helpful to develop self-regulation skills in attention training through a personalized game experience. It can provide children with appropriate guidance to develop their self-regulation skills while enhancing their cognitive functions such as attention. The transparency of the user model may help users (both children and tutors) to follow their progression, and to develop their learning strategy. In this paper, a pilot study was conducted to evaluate the assessment step of a developed attention training serious game. The objective was to study whether the transparency of the learner model influences the users' perception of their attention and self-regulate their learning. The primary results of the pilot experiment show that open learner model influences the decision of users on difficulty level preference that may be promising in self-regulated attention training.
The increasing amount of data in social media enables new advanced user modeling approaches. This paper focuses on user profiling for diversity-enhancing recommender systems for finding new followees on Twitter. By combining social network analysis with Latent Dirichlet Allocation based content analysis, we defined three egocentric structural positions on the network extracted from Twitter data: Mentions of Mentions, Community Cluster, Dormant Ties (and the rest as a baseline condition). In addition to describing the data analysis procedure, we report preliminary empirical findings on a user-centered evaluation study of recommendations based on the proposed matching strategy and the presented structural positions. The investigation of the possible overlaps of the groups and the participants' evaluations of perceived relevance of the recommendation imply that the three positions are sufficiently mutually exclusive and thus could serve as new diversity-enhancing mechanisms in various people recommender systems.
In this paper we point out some relevant issues in relation to privacy when providing holistic recommendations. We emphasize that a holistic recommender should be fair, explainable and privacy-preserving to ensure the ethicality of the recommendation process. Further, we point out relevant research questions that should be addressed in the future, as well as propose some preliminary suggestions to face the emergent issues with reference to privacy in the recommendation domain.
As already pointed out by a constantly growing literature, explainability in recommender systems field is a key aspect to increase users' satisfaction. With the increase of user generated content, tags have proven to be highly relevant when it comes to describe either users or items. A number of strategies that rely on tags have been proposed, yet, many of these algorithms exploit the frequency of user-tags interactions to gain information. We argue that a pure frequentist description might lack of specificity to grasp user's peculiar tastes. Therefore, we propose a novel approach based on game theory that tries to find the best trade-off between generality and detailing. The identified user's description can be used to keep her in the loop and allows the user to have control over system's knowledge. Additionally, we propose a user interface that embeds the proposed user's description and it can be used by the user herself to guide her catalogue's exploration toward novel and serendipitous items.
Students might pursue different goals throughout their learning process. For example, they might be seeking new material to expand their current level of knowledge, repeating content of prior classes to prepare for an exam, or working on addressing their most recent misconceptions. Multiple potential goals require an adaptive e-learning system to recommend learning content appropriate for students' intent and to explain this recommendation in the context of this goal. In our prior work, we explored explainable recommendations for the most typical 'knowledge expansion goal". In this paper, we focus on students' immediate needs to remedy misunderstandings when they solve programming problems. We generate learning content recommendations to target the concepts with which students have struggled more recently. At the same time, we produce explanations for this recommendation goal in order to support students' understanding of why certain learning activities are recommended. The paper provides an overview of the design of this explainable educational recommender system and describes its ongoing evaluation
SESSION: Fairness in User Modeling, Adaptation and Personalization (FairUMAP 2019) Workshop
It is our great pleasure to welcome you to the Second FairUMAP workshop at UMAP 2019. This full-day workshop brings together researchers working at the intersection of user modeling, adaptation, and personalization on one hand, and bias, fairness and transparency in algorithmic systems on the other hand. The workshop was motivated by the observation that these two fields increasingly impact one another. Personalization has become a ubiquitous and essential part of systems that help users find relevant information in today's highly complex, information-rich online environments. Machine learning techniques applied to big data, as done by recommender systems, and user modeling in general, are key enabling technologies that allow intelligent systems to learn from users and adapt their output to users' needs and preferences. However, there has been a growing recognition that these underlying technologies raise novel ethical, legal, and policy challenges. It has become apparent that a single-minded focus on user characteristics has obscured other important and beneficial outcomes such systems must be able to deliver. System properties such as fairness, transparency, balance, and other social welfare considerations are not captured by typical metrics based on which data-driven personalized models are optimized. Indeed, widely-used personalization systems in popular sites such as Facebook, Google News and YouTube have been heavily criticized for personalizing information delivery too heavily at the cost of these other objectives.
Recommender systems (RS) have been introduced to educations as an effective technology-enhanced learning technique. Traditional RS produce recommendations by considering the preferences of the end users only. Multi-stakeholder recommender systems (MSRS) claim that it is necessary to consider the utility of the items from the perspective of other stakeholders in order to balance the needs of multiple stakeholders. Take book recommendations for example, the utility of items from the view of parents, instructors and even publishers may be also important in addition to the student preferences. In this paper, we propose and exploit utility-based MSRS for personalized learning. Particularly, we attempt to address the challenge of over-/under-expectations in the utility-based MSRS. Our experimental results based on an educational data demonstrate the effectiveness of our proposed models and solutions.
Research on user modeling and personalization typically only serves the needs of end-users. However, when applied in real-world, commercial contexts, recommendations should also serve the (often monetary) interests of other parties, such as platform providers, sellers and advertisers. This paper provides a brief historical perspective on the research field, contrasts this with the commercial context, and investigates the topics currently addressed at the UMAP and RecSys conferences. The paper concludes with a discussion on the need for the research community to take multi-stakeholder interests into account in the design and evaluation of adaptive systems. This would allow us to foresee unwanted effects, such as online filter bubbles, and to pro-actively find strategies to prevent them.
Recent research in fairness in machine learning has identified situations in which biases in input data can cause harmful or unwanted effects. Researchers in the areas of personalization and recommendation have begun to study similar types of bias. What these lines of research share is a fixed representation of the protected groups relative to which bias must be monitored. However, in some real-world application contexts, such groups cannot be defined apriori, but must be derived from the data itself. Furthermore, as we show, it may be insufficient in such cases to examine global system properties to identify protected groups. Thus, we demonstrate that fairness may be local, and the identification of protected groups only possible through consideration of local conditions.
User modeling has become an indispensable feature of a plethora of different digital services such as search engines, social media or e-commerce. Indeed, decision procedures of online algorithmic systems apply various methods including machine learning (ML) to generate virtual models of billions of human beings based on large amounts of personal and other data. Recently, there has been a call for a "Right to Reasonable Inferences" for Europe's General Data Protection Regulation (GDPR). Here, we explore a conceptualization of reasonable inference in the context of image analytics that refers to the notion of evidence in theoretical reasoning. The main goal of this paper is to start defining principles for reasonable image inferences, in particular, portraits of individuals. Based on an image analytics case study, we use the notions of first- and second-order inferences to determine the reasonableness of predicted concepts. Finally, we highlight three key challenges for the future of this research space: first, we argue for the potential value of hidden quasi-semantics. Second, we indicate that automatic inferences can create a fundamental trade-off between privacy preservation and "model fit" and, third, we end with the question whether human reasoning can serve as a normative benchmark for reasonable automatic inferences.
In this work, we investigate whether privacy and fairness can be simultaneously achieved by a single classifier in several different models. Some of the earliest work on fairness in algorithm design defined fairness as a guarantee of similar outputs for "similar'' input data, a notion with tight technical connections to differential privacy. We study whether tensions exist between differential privacy and statistical notions of fairness, namely Equality of False Positives and Equality of False Negatives (EFP/EFN). We show that even under full distributional access, there are cases where the constraint of differential privacy precludes exact EFP/EFN. We then turn to ask whether one can learn a differentially private classifier which approximately satisfies EFP/EFN, and show the existence of a PAC learner which is private and approximately fair with high probability. We conclude by giving an efficient algorithm for classification that maintains utility and satisfies both privacy and approximate fairness with high probability.
SESSION: Human Aspects in Adaptive and Personalized Interactive Environments (HAAPIE 2019) Workshop
UMAP 2019 HAAPIE (Human Aspects in Adaptive and Personalized Interactive Environments) Workshop Chairs' Welcome
It is our great pleasure to welcome you to the 4th International Workshop on Human Aspects in Adaptive and Personalized Interactive Environments (HAAPIE 2019). HAAPIE 2019 (http://haapie.cs.ucy.ac.cy) is a full-day workshop held on 09 June 2019 in conjunction with the 27th ACM Conference on User Modeling, Adaptation and Personalization (UMAP 2019), 09-12 June 2019 in Larnaca, Cyprus. Nowadays, the profound digital transformation has upgraded the role of the computational system into an intelligent multidimensional communication medium that creates new opportunities, competencies, models and processes. HAAPIE embraces the essence of the human-machine co-existence and aims to bring more inclusively the 'human-in-the-loop" approach/idea, adequately supporting the rising multi-purpose goals, needs, requirements, activities and interactions of users through new human-centered adaptive and personalized interactive environments, algorithms and systems. It brings together experts, researchers, students and practitioners from different disciplines in order to share ideas and experiences, lessons learned, approaches and results that could substantially contribute to the broader UMAP community. This year we received 10 submissions from all around the world covering a broad range of topics on the workshop's research themes. Each paper has been reviewed by up to 3 members of the IPC with expertise in the respective area to ensure the necessary relevance, quality and novelty.
Personalization in principle cannot happen without information about individuals, requiring personalization systems to comply with official privacy regulations. However, in order to design personalization systems that provide the best possible privacy-related user experience, a more human-centered perspective has to be taken into account. As a first step towards this goal, in the present work we show the setup and results of an online survey investigating the relation between the intention to disclose certain categories of personal data and the type of benefit promised by personalization.
Shaping the Reaction: Community Characteristics and Emotional Tone of Citizen Responses to Robotics Videos at TED versus YouTube
When modelling for the social we need to consider more than one medium. Little is known as to how platform community characteristics shape the discussion and how communicators could best engage each community, taking into consideration these characteristics. We consider comments on TED videos featuring roboticists, shared at TED.com and YouTube. We find evidence of different social norms and importantly, approaches to comment writing. The emotional tone is more positive at TED; however, there is little emotional escalation in either platform. The study highlights the importance of considering the community characteristics of a medium, when communicating with the public in a case study of emerging technologies.
A critical phase in teaching is the effective design of educational contents. Instructors are phased with the dilemma of compensating on the volume and complexity that academic curriculum may entail, to easily accommodating educational content to learners. InfoVis or Infographics feature as a viable method to alleviate this problem through rich information and structured visual stories by taking advantage of the visual thinking of individuals and difficulties in information processing. A challenge, however, is the creation of personalized educational content that will dynamically adapt to users' intrinsic cognitive and emotional characteristics. We present a preliminary user study that explores two human factors, i.e., metacognition and motivation, which could enrich user models and guide the personalization process of learning material devised as infographic. Our results revealed strong influence of the two human factors in the learning process, while in cases suggest that may also be used as good predictors of academic achievement.
Personalization is typically based on preferences extracted from the interactions of users with the system. A recent trend is to also account for the social influence among users, which may play a non-negligible role in shaping one's individual preferences. The underlying assumptions are that friends tend to develop similar taste, i.e., homophily, and that similar users tend to connect to each other, i.e., social selection. In this work, we investigate the conditions under which social influence has a significant impact on the preferences of users. We find that pairs of friends, where one is socially very active whereas the other is not, exhibit stronger correlations in their preferences compared to other pairs of friends, implying thus a stronger mechanism of influence.
User reviews of apps are critically important in open mobile application markets, including the App Store and Google Play. Analyzing app reviews helps reveal any usability issues faced, desired improvements, and could also provide insights to guide future app designs. As a result, there is a growing demand for analysis of app reviews to enhance app usability, user experience, and hence improve overall app adoption. This is particularly true for apps targeting sensitive issues such as those promoting mental health. In this paper, we present the results of an analysis of 106 mental health app reviews from the App Store and Google Play. We mined and analyzed 1236 distinct reviews to identify usability issues. We classified app usability issues into six categories: bugs, poor user interface design, data loss, battery and memory usage issue, lack of guidance and explanation, and internet connectivity issue. The results could guide app designers on how to design apps especially those tailored to mental health to improve their usability.
Menopause is a natural part of women's aging, but is often accompanied by an increased cardiometabolic risk (CMR), of which most women are unaware. Preventive self-care via mobile health applications (apps) is a promising way to address this issue, but research on apps for middle-aged women is limited. Further, modeling such risk is no trivial task in a non-clinical self-care context, where most biomarkers used in traditional models are unavailable. Machine learning (ML) is a potential option in this regard, but many ML approaches are effectively black box models, which leads to doubt regarding their trustworthiness. Therefore, in this paper we analyze and compare different decision tree and rule-based classification models, considered to be inherently interpretable, to assess the CMR of early middle-aged women in the context of a non-clinical self-care app. For this, we first defined a set of candidate determinants based on the feedback of potential users and domain experts. We then used data from a subset of the participants in the Study of Women's Health Across the Nation (SWAN) to compare these ML models with traditional risk score models, based on five cardiometabolic 10-year outcomes: heart attack, stroke, angina pectoris, diabetes, and metabolic syndrome.
Virtual labs enable inquiry-based learning where students can implement their own experiments using virtual objects and apparatus. Although the benefits of adaptive and personalised learning are well recognised, these were not thoroughly investigated in virtual labs. This paper presents the architecture of an interactive science virtual lab that personalises the learning journey based on the student's self-directed learning (SDL) and self-efficacy (SE) levels. The results of a pilot in two secondary schools showed that both students with low and high SDL and SE level improved their knowledge, but students with low SDL and SE had a higher number of incorrect attempts before completing the experiment.
SESSION: Personalized access to Cultural Heritage (PATCH 2019) Workshop
It is our great pleasure to welcome you to the ACM 2019 PATCH. Following the successful series of PATCH workshops, started in 2007, PATCH 2019 is organized as the meeting point between state of the art cultural heritage research and personalization - using any kind of technology, while focusing on ubiquitous and adaptive scenarios, to enhance the personal experience in cultural heritage sites. The workshop is aimed at bringing together researchers and practitioners who are working on various aspects of cultural heritage and are interested in exploring the potential of state of the art of personalized approaches that may enhance the CH visit experience.
In this paper, we present the results of a user study focusing on whether the visitors' cultural preferences and expectations relate to different personality traits as defined by the Big Five personality model. We describe the user study procedure and report the correlations discovered between some of the Big Five factors and the participant assessments, over particular aspects of a shared digital storytelling experience. We suggest that the results may notably inform not only the design but also the evaluation of cultural experiences, laying the foundations for a promising line of work.
Thematic maps, traditionally developed to present specific themes within defined geographical areas, are an interesting information presentation model for Cultural Heritage exploration because of the abstract view on the territory they provide. However, in order to cope with possibly heterogeneous user interests, they should be adapted to the individual user by including the relevant types of information, given her/his specific interests. In a previous paper, we proposed an approach to the integration of thematic maps in the OnToMap Participatory GIS (Geographic Information System), in order to support query expansion during an exploratory search task. The proposed maps were built on the basis of a survey in which we asked people to rate the relevance of a set of concepts to five main themes around which we developed the maps. In this paper we go one step forward and we propose a more general approach to information search support in order to automatically create thematic maps, based on the analysis of frequently co-occurring search interests in a search engine query log. This type of analysis supports the identification of clusters of concepts that people frequently search within the same sessions and helps the identification of co-occurring topics that can be proposed to users when exploring an information space. In this way, when the user browses a catalog of Cultural Heritage information, (s)he can both visualize the thematic maps relevant to the search context, and be guided in the navigation within types of information, looking for possibly complementary types of data to satisfy her/his needs.
In this digital era, one of the main challenge faced by cultural heritage is digitization. This challenge is particularly hard in countries like Italy, characterized by an extremely high number of Cultural goods. Data acquisition for many of these Cultural Heritage is extremely difficult, because of the complexity of surveys through traditional methodologies. In this paper, we propose a novel approach to the knowledge and data acquisition Cultural Heritage based on social media. The proposed approach, named "HeritageGo" (HeGo), transforms the user as an actor of the procedures for the acquisition of raw data. The paper also describes the first experiments focusing on the metric quality of the models obtained with SfM methodologies from raw data acquired by users.
This paper describes an exploratory study that attempts to classify museum visitors by taking into consideration indoor behavior and demographic features. We discuss different approaches of using such data for improving the user experience in the museum. Moreover, we try to explain user's behavior by creating different user groups using a novel data set. Our findings indicate that knowing user age, education and her museum visits frequency, together with the current visit signals (total standing time and listening to a mobile guide time) can be used for visitors classification that might be useful in designing new intelligent user interfaces that can improve the visitor's indoor experience.
Brain Computer Interfaces (BCIs) are interfaces that put the user in communication with an electronic device through the brain activity produced by the user herself. Non-invasive BCI are mainly based on electroencephalographic (EEG) signals. While using these systems, users become able to manipulate their brain activity to produce signals that will then be used to control computers or Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than ACM must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Request permissions from firstname.lastname@example.org. UMAP '19, June 9-12, 2019, Larnaca, Cyprus © 2019 Association for Computing Machinery. ACM ISBN 978-1-4503-6021-0/19/06. . . $15.00 https://doi.org/10.1145/3320435.3320472 other devices without the aid of motor movements. Besides active BCI systems, in which the user directly control the system by a conscious and voluntary mental activity, there are passive BCI that can be used to recognize mental states, like the user's emotional state (in particular the level of engagement) during the interaction or according to a received stimulus. In this paper we describe a passive BCI experiment and its results, wherein users are exposed to a set of emotional artworks and the engagement measured through a BCI headset is compared to their explicit engagement, in order to test the reliability of BCI-based engagement detection
Investigating the multimodal communication of Tourist Guides to implement a Virtual Tourist Guide leading tourists in three Italian Charterhouses, the paper focuses on an aspect of the human guide's speech that would be useful to create a very realistic Virtual Guide: linguistic disfluencies. On a corpus of three guided tours in S. Martino Charterhouse (Naples) an analysis is presented of the guides' pauses and their concomitant gestures.
It is widely agreed that museums and other cultural heritage venues should provide visitors with personalised interaction and services such as personalised mobile guides, although currently most do not. Since museum visitors are typically first-time visitors and since their visit is for a relatively short session, personalisation should use initial interaction data to associate the user with a particular persona and thereby infer other facts about the user's preferences and needs. In this paper we report a questionnaire-based study carried out with 105 visitors of a Science and Technology Centre to examine the minimal features needed to identify visitor personas. We find that museum visitors can be clustered by their visit motivation and perceived success factors; these clusters are found to correspond both with Falk's visitor categorisation and a prior classification of exploration styles. Consequently, these two features can be used to reliably identify the visitor persona, and therefore, can be used for user modeling.
The heterogeneity of the audience of cultural heritage (CH) institutions introduces numerous challenges to the delivery of meaningful CH content. People with diverse cognitive characteristics are engaged with varying CH activities (e.g., games, guided visits) and considering that cognitive characteristics define the way we process information, our experience, behavior, and knowledge acquisition are influenced. Our recent studies provide evidence that human cognition should be considered as an important personalization factor within CH contexts, and thus, we developed a framework that delivers cognition-centered personalized CH activities. The efficiency and the efficacy of the framework have been successfully assessed, but, non-technical users may face difficulties when attempting to use it and create personalized CH activities. In this paper, we present DeCACHe which supports CH designers in developing cognition-centered personalized CH activities throughout different aspects of the design lifecycle. We also report a case study with one CH designer, who used our tool to produce two versions of his CH game for people with different cognitive characteristics.
In this paper, we describe our research activities for integrating the recommendation process of nearby points of artistic and cultural interest (POIs) with related multimedia content. The recommendation engine exploits the potential offered by linked open data (LOD), by following semantic links in the LOD graph to identify movies, books, and music artists/songs related to that specific POI. This content is subsequently reranked based on the activity of the user and her friends on social media (i.e., Facebook), in order to provide personalized suggestions.
Investigating the Effectiveness of Narrative Relations for the Exploration of Cultural Heritage Archives: A Case Study on the Labyrinth system
Exploratory environments can overcome the limitations of keyword-based access, criticized for the difficulty of formulating specificqueries in large and complex domains and for its incapability ofpromoting personalized exploration paths. In this paper, we describe a task-based user study (n=16) aimed at testing the use ofa narrative model for supporting exploratory search in culturalheritage archives. The results show that a narrative model can actually support the exploration of cultural heritage archives, accommodating unknown and unexpected elements and relationsinto user-created paths, although further investigation on failed tasks is needed.
This paper is a first step towards identifying the links between the characteristics of gaze behaviour and visitor preferences in a museum. In the long term, the real-time analysis of visitors' gaze should allow a fine estimation of their interest for the different artworks exhibited and should replace the fastidious and time-consuming elicitation of preferences commonly used in traditional recommender systems. To study these links, we carried out a user study at the Nancy Museum of Fine Arts in the North-East of France. This pilot study involved 13 volunteers who had the opportunity to freely explore the museum and contemplate hundreds of artworks for more than 50 minutes on average in May 2018. We were able to analyze millions of fixation points so as to find correlations between the number of fixation points per painting, the time spent looking at a painting, and whether or not this painting is appreciated. We plan to extend this study to 100 visitors in the coming months.
SESSION: ACM UMAP 2019 Tutorials
This tutorial will introduce User Modeling (UM) researchers to the techniques of empirical evaluation of user modeling systems. No background in statistics is required. The target audience is UM researchers, especially students, who have a background in computer science or some other field that does not normally include designing and running human-subject experiments. Topics include designing experiments (choosing independent/dependent variables, covariant and nuisance variables, between vs. within subjects designs, factorial designs, estimating sensitivity, layered evaluation), running experiments (recruiting participants, controlling the environment, recording data), data analysis (statistical tests, ANOVA, checking assumptions of statistical methods, multiple testing correction, explained variance), and common surveys/tests for gathering covariate data.
Distributed Ledger (blockchain) technology provides an alternative for distributed databases. It creates a secure and immutable record of data transactions, thus tracking how data are shared and accessed. The access and operations on data can be regulated via "smart contracts" that allows setting conditions for accessing the data - by whom, for what purpose, for how long, under what conditions, whether access is granted to the original data or just a query /derivative data. In addition, users can benefit from sharing their data by using "smart contracts" that regulate sharing for monetary reward, or for another form of recognition. Both user profile data and user-owned data can be shared in this way, empowering users to benefit from their data, under their own conditions, rather than surrendering it to centralized services. The tutorial will present the basics of distributed ledger technology and smart contracts and will train the participants in using a privacy-preserving user data-sharing framework.