Showing papers on "User modeling published in 2015"

A Multi-View Deep Learning Approach for Cross Domain User Modeling in Recommendation Systems

Abstract: Recent online services rely heavily on automatic personalization to recommend relevant content to a large number of users. This requires systems to scale promptly to accommodate the stream of new users visiting the online services for the first time. In this work, we propose a content-based recommendation system to address both the recommendation quality and the system scalability. We propose to use a rich feature set to represent users, according to their web browsing history and search queries. We use a Deep Learning approach to map users and items to a latent space where the similarity between users and their preferred items is maximized. We extend the model to jointly learn from features of items from different domains and user features by introducing a multi-view Deep Learning model. We show how to make this rich-feature based user representation scalable by reducing the dimension of the inputs and the amount of training data. The rich user feature representation allows the model to learn relevant user behavior patterns and give useful recommendations for users who do not have any interaction with the service, given that they have adequate search and browsing history. The combination of different domains into a single model for learning helps improve the recommendation quality across all the domains, as well as having a more compact and a semantically richer user latent feature vector. We experiment with our approach on three real-world recommendation systems acquired from different sources of Microsoft products: Windows Apps recommendation, News recommendation, and Movie/TV recommendation. Results indicate that our approach is significantly better than the state-of-the-art algorithms (up to 49% enhancement on existing users and 115% enhancement on new users). In addition, experiments on a publicly open data set also indicate the superiority of our method in comparison with transitional generative topic models, for modeling cross-domain recommender systems. Scalability analysis show that our multi-view DNN model can easily scale to encompass millions of users and billions of item entries. Experimental results also confirm that combining features from all domains produces much better performance than building separate models for each domain.

Recommender systems based on user reviews: the state of the art

Abstract: In recent years, a variety of review-based recommender systems have been developed, with the goal of incorporating the valuable information in user-generated textual reviews into the user modeling and recommending process. Advanced text analysis and opinion mining techniques enable the extraction of various types of review elements, such as the discussed topics, the multi-faceted nature of opinions, contextual information, comparative opinions, and reviewers' emotions. In this article, we provide a comprehensive overview of how the review elements have been exploited to improve standard content-based recommending, collaborative filtering, and preference-based product ranking techniques. The review-based recommender system's ability to alleviate the well-known rating sparsity and cold-start problems is emphasized. This survey classifies state-of-the-art studies into two principal branches: review-based user profile building and review-based product profile building. In the user profile sub-branch, the reviews are not only used to create term-based profiles, but also to infer or enhance ratings. Multi-faceted opinions can further be exploited to derive the weight/value preferences that users place on particular features. In another sub-branch, the product profile can be enriched with feature opinions or comparative opinions to better reflect its assessment quality. The merit of each branch of work is discussed in terms of both algorithm development and the way in which the proposed algorithms are evaluated. In addition, we discuss several future trends based on the survey, which may inspire investigators to pursue additional studies in this area.

Click Models for Web Search

Abstract: With the rapid growth of web search in recent years the problem of modeling its users has started to attract more and more attention of the information retrieval community. This has several motivations. By building a model of user behavior we are essentially developing a better understanding of a user, which ultimately helps us to deliver a better search experience. A model of user behavior can also be used as a predictive device for non-observed items such as document relevance, which makes it useful for improving search result ranking. Finally, in many situations experimenting with real users is just infeasible and hence user simulations based on accurate models play an essential role in understanding the implications of algorithmic changes to search engine results or presentation changes to the search engine result page. In this survey we summarize advances in modeling user click behavior on a web search engine result page. We present simple click models as well as more complex models aimed at capturing non-trivial user behavior patterns on modern search engine result pages. We discuss how these models compare to each other, what challenges they have, and what ways there are to address these challenges. We also study the problem of evaluating click models and discuss the main applications of click models.

Evaluating Recommender Systems

Abstract: Recommender systems are now popular both commercially and in the research community, where many approaches have been suggested for providing recommendations. In many cases a system designer that wishes to employ a recommendater system must choose between a set of candidate approaches. A first step towards selecting an appropriate algorithm is to decide which properties of the application to focus upon when making this choice. Indeed, recommender systems have a variety of properties that may affect user experience, such as accuracy, robustness, scalability, and so forth. In this paper we discuss how to compare recommenders based on a set of properties that are relevant for the application. We focus on comparative studies, where a few algorithms are compared using some evaluation metric, rather than absolute benchmarking of algorithms. We describe experimental settings appropriate for making choices between algorithms. We review three types of experiments, starting with an offline setting, where recommendation approaches are compared without user interaction, then reviewing user studies, where a small group of subjects experiment with the system and report on the experience, and finally describe large scale online experiments, where real user populations interact with the system. In each of these cases we describe types of questions that can be answered, and suggest protocols for experimentation. We also discuss how to draw trustworthy conclusions from the conducted experiments. We then review a large set of properties, and explain how to evaluate systems given relevant properties. We also survey a large set of evaluation metrics in the context of the property that they evaluate.

Expert Finding for Question Answering via Graph Regularized Matrix Completion

Abstract: Expert finding for question answering is a challenging problem in community-based question answering (CQA) systems, arising in many real applications such as question routing and identification of best answers. In order to provide high-quality experts, many existing approaches learn the user model from their past question-answering activities in CQA systems. However, the past activities of users in most CQA systems are rather few, and thus the user model may not be well inferred in practice. In this paper, we consider the problem of expert finding from the viewpoint of missing value estimation. We then employ users’ social networks for inferring user model, and thus improve the performance of expert finding in CQA systems. In addition, we develop a novel graph-regularized matrix completion algorithm for inferring the user model. We further develop two efficient iterative procedures, GRMC-EGM and GRMC-AGM, to solve the optimization problem. GRMC-EGM utilizes the Extended Gradient Method (EGM), while GRMC-AGM applies the Accelerated proximal Gradient search Method (AGM), for the optimization. We evaluate our methods on the well-known question answering system Quora, and the popular social network Twitter. Our empirical study shows the effectiveness of the proposed algorithms in comparison to the state-of-the-art expert finding algorithms.

Dynamic User Modeling in Social Media Systems

Abstract: Social media provides valuable resources to analyze user behaviors and capture user preferences. This article focuses on analyzing user behaviors in social media systems and designing a latent class statistical mixture model, named temporal context-aware mixture model (TCAM), to account for the intentions and preferences behind user behaviors. Based on the observation that the behaviors of a user in social media systems are generally influenced by intrinsic interest as well as the temporal context (e.g., the public's attention at that time), TCAM simultaneously models the topics related to users' intrinsic interests and the topics related to temporal context and then combines the influences from the two factors to model user behaviors in a unified way. Considering that users' interests are not always stable and may change over time, we extend TCAM to a dynamic temporal context-aware mixture model (DTCAM) to capture users' changing interests. To alleviate the problem of data sparsity, we exploit the social and temporal correlation information by integrating a social-temporal regularization framework into the DTCAM model. To further improve the performance of our proposed models (TCAM and DTCAM), an item-weighting scheme is proposed to enable them to favor items that better represent topics related to user interests and topics related to temporal context, respectively. Based on our proposed models, we design a temporal context-aware recommender system (TCARS). To speed up the process of producing the top-k recommendations from large-scale social media data, we develop an efficient query-processing technique to support TCARS. Extensive experiments have been conducted to evaluate the performance of our models on four real-world datasets crawled from different social media sites. The experimental results demonstrate the superiority of our models, compared with the state-of-the-art competitor methods, by modeling user behaviors more precisely and making more effective and efficient recommendations.

User modeling with neural network for review rating prediction

Abstract: We present a neural network method for review rating prediction in this paper. Existing neural network methods for sentiment prediction typically only capture the semantics of texts, but ignore the user who expresses the sentiment. This is not desirable for review rating prediction as each user has an influence on how to interpret the textual content of a review. For example, the same word (e.g. "good") might indicate different sentiment strengths when written by different users. We address this issue by developing a new neural network that takes user information into account. The intuition is to factor in user-specific modification to the meaning of a certain word. Specifically, we extend the lexical semantic composition models and introduce a userword composition vector model (UWCVM), which effectively captures how user acts as a function affecting the continuous word representation. We integrate UWCVM into a supervised learning framework for review rating prediction, and conduct experiments on two benchmark review datasets. Experimental results demonstrate the effectiveness of our method. It shows superior performances over several strong baseline methods.

Understanding Your Users: A Practical Guide to User Research Methods

Abstract: This new and completely updated edition is a comprehensive, easy-to-read, "how-to" guide on user research methods. You'll learn about many distinct user research methods and also pre- and post-method considerations such as recruiting, facilitating activities or moderating, negotiating with product developments teams/customers, and getting your results incorporated into the product. For each method, you'll understand how to prepare for and conduct the activity, as well as analyze and present the data - all in a practical and hands-on way. Each method presented provides different information about the users and their requirements (e.g., functional requirements, information architecture). The techniques can be used together to form a complete picture of the users' needs or they can be used separately throughout the product development lifecycle to address specific product questions. These techniques have helped product teams understand the value of user experience research by providing insight into how users behave and what they need to be successful. You will find brand new case studies from leaders in industry and academia that demonstrate each method in action. This book has something to offer whether you are new to user experience or a seasoned UX professional. After reading this book, you'll be able to choose the right user research method for your research question and conduct a user research study. Then, you will be able to apply your findings to your own products. Completely new and revised edition includes 30+% new content! Discover the foundation you need to prepare for any user research activity and ensure that the results are incorporated into your products Includes all new case studies for each method from leaders in industry and academia Table of Contents Preface Part 1: What You Need to Know Before Choosing An Activity 1. Introduction to User Experience 2. Before You Choose an Activity: Learning About Your Product and Users 3. Ethical and Legal Considerations 4. Setting Up Research Facilities 5. Choosing a User Experience Research Activity Part 2: Get Up and Running 6. Preparing For Your User Research Activity 7. During Your User Research Activity Part 3: The Methods 8. Diary Studies 9. Interviews 10. Surveys 11. Card Sorting 12. Focus Groups 13. Field Studies 14. Evaluating Methods Part 4: Wrapping Up 15. Concluding Final Part 5: Appendices Appendix A. Requirements for Creating a Participant Recruitment Database Appendix B. Report Template Appendix C. Glossary Appendix D. References Index

User-controllable personalization

Abstract: In this research we investigated the role of user controllability on personalized systems by implementing and studying a novel interactive recommender interface, SetFusion. We examined whether allowing the user to control the process of fusing or integrating different algorithms (i.e., different sources of relevance) resulted in increased engagement and a better user experience. The essential contribution of this research stems from the results of a user study (N=40) of controllability in a scenario where users could fuse different recommendation approaches, with the possibility of inspecting and filtering the items recommended. First, we introduce an interactive Venn diagram visualization, which combined with sliders, can provide an efficient visual paradigm for information filtering. Second, we provide a three-fold evaluation of the user experience: objective metrics, subjective user perception, and behavioral measures. Through the analysis of these metrics, we confirmed results from recent studies, such as the effect of trusting propensity on accepting the recommendations and also unveiled the importance of features such as being a native speaker. Our results present several implications for the design and implementation of user-controllable personalized systems. We explored user-controllable interfaces as extension of traditional-ranked lists.We introduced SetFusion, a controllable interface with sliders and a Venn diagram.We conducted a controlled user study on online conference article recommendation.Our evaluation had three dimensions: users' perception, behavioral and IR metrics.Controllable interface had a positive effect influenced by users' characteristics.

User Conditional Hashtag Prediction for Images

Abstract: Understanding the content of user's image posts is a particularly interesting problem in social networks and web settings. Current machine learning techniques focus mostly on curated training sets of image-label pairs, and perform image classification given the pixels within the image. In this work we instead leverage the wealth of information available from users: firstly, we employ user hashtags to capture the description of image content; and secondly, we make use of valuable contextual information about the user. We show how user metadata (age, gender, etc.) combined with image features derived from a convolutional neural network can be used to perform hashtag prediction. We explore two ways of combining these heterogeneous features into a learning framework: (i) simple concatenation; and (ii) a 3-way multiplicative gating, where the image model is conditioned on the user metadata. We apply these models to a large dataset of de-identified Facebook posts and demonstrate that modeling the user can significantly improve the tag prediction quality over current state-of-the-art methods.

User Activity Patterns During Information Search

Abstract: Personalization of support for information seeking depends crucially on the information retrieval system's knowledge of the task that led the person to engage in information seeking. Users work during information search sessions to satisfy their task goals, and their activity is not random. To what degree are there patterns in the user activity during information search sessionsq Do activity patterns reflect the user's situation as the user moves through the search task under the influence of his or her task goalq Do these patterns reflect aspects of different types of information-seeking tasksq Could such activity patterns identify contexts within which information seeking takes placeq To investigate these questions, we model sequences of user behaviors in two independent user studies of information search sessions (N = 32 users, 128 sessions, and N = 40 users, 160 sessions). Two representations of user activity patterns are used. One is based on the sequences of page use; the other is based on a cognitive representation of information acquisition derived from eye movement patterns in service of the reading process. One of the user studies considered journalism work tasks; the other concerned background research in genomics using search tasks taken from the TREC Genomics Track. The search tasks differed in basic dimensions of complexity, specificity, and the type of information product (intellectual or factual) needed to achieve the overall task goal. The results show that similar patterns of user activity are observed at both the cognitive and page use levels. The activity patterns at both representation layers are able to distinguish between task types in similar ways and, to some degree, between tasks of different levels of difficulty. We explore relationships between the results and task difficulty and discuss the use of activity patterns to explore events within a search session. User activity patterns can be at least partially observed in server-side search logs. A focus on patterns of user activity sequences may contribute to the development of information systems that better personalize the user's search experience.

A Comparison of Offline Evaluations, Online Evaluations, and User Studies in the Context of Research-Paper Recommender Systems

Abstract: The evaluation of recommender systems is key to the successful application of recommender systems in practice. However, recommender-systems evaluation has received too little attention in the recommender-system community, in particular in the community of research-paper recommender systems. In this paper, we examine and discuss the appropriateness of different evaluation methods, i.e. offline evaluations, online evaluations, and user studies, in the context of research-paper recommender systems. We implemented different content-based filtering approaches in the research-paper recommender system of Docear. The approaches differed by the features to utilize (terms or citations), by user model size, whether stop-words were removed, and several other factors. The evaluations show that results from offline evaluations sometimes contradict results from online evaluations and user studies. We discuss potential reasons for the non-predictive power of offline evaluations, and discuss whether results of offline evaluations might have some inherent value. In the latter case, results of offline evaluations were worth to be published, even if they contradict results of user studies and online evaluations. However, although offline evaluations theoretically might have some inherent value, we conclude that in practice, offline evaluations are probably not suitable to evaluate recommender systems, particularly in the domain of research paper recommendations. We further analyze and discuss the appropriateness of several online evaluation metrics such as click-through rate, link-through rate, and cite-through rate.

The Retrieval Effectiveness of Web Search Engines: Considering Results Descriptions

Abstract: Purpose: To compare five major Web search engines (Google, Yahoo, MSN, Ask.com, and Seekport) for their retrieval effectiveness, taking into account not only the results but also the results descriptions. Design/Methodology/Approach: The study uses real-life queries. Results are made anonymous and are randomised. Results are judged by the persons posing the original queries. Findings: The two major search engines, Google and Yahoo, perform best, and there are no significant differences between them. Google delivers significantly more relevant result descriptions than any other search engine. This could be one reason for users perceiving this engine as superior. Research Limitations: The study is based on a user model where the user takes into account a certain amount of results rather systematically. This may not be the case in real life. Practical Implications: Implies that search engines should focus on relevant descriptions. Searchers are advised to use other search engines in addition to Google. Originality/Value: This is the first major study comparing results and descriptions systematically and proposes new retrieval measures to take into account results descriptions

User Interaction Models for Disambiguation in Programming by Example

Abstract: Programming by Examples (PBE) has the potential to revolutionize end-user programming by enabling end users, most of whom are non-programmers, to create small scripts for automating repetitive tasks. However, examples, though often easy to provide, are an ambiguous specification of the user's intent. Because of that, a key impedance in adoption of PBE systems is the lack of user confidence in the correctness of the program that was synthesized by the system. We present two novel user interaction models that communicate actionable information to the user to help resolve ambiguity in the examples. One of these models allows the user to effectively navigate between the huge set of programs that are consistent with the examples provided by the user. The other model uses active learning to ask directed example-based questions to the user on the test input data over which the user intends to run the synthesized program. Our user studies show that each of these models significantly reduces the number of errors in the performed task without any difference in completion time. Moreover, both models are perceived as useful, and the proactive active-learning based model has a slightly higher preference regarding the users' confidence in the result.

Unified YouTube Video Recommendation via Cross-network Collaboration

Abstract: The ever growing number of videos on YouTube makes recommendation an important way to help users explore interesting videos. Similar to general recommender systems, YouTube video recommendation suffers from typical problems like new user, cold-start, data sparsity, etc. In this paper, we propose a unified YouTube video recommendation solution via cross-network collaboration: users' auxiliary information on Twitter are exploited to address the typical problems in single network-based recommendation solutions. The proposed two-stage solution first transfers user preferences from auxiliary network by learning cross-network behavior correlations, and then integrates the transferred preferences with the observed behaviors on target network in an adaptive fashion. Experimental results show that the proposed cross-network collaborative solution achieves superior performance not only in term of accuracy, but also in improving the diversity and novelty of the recommended videos.

Context based customization of word assistance functions

Abstract: For generating customized word assistance functions based on user information and context, a system, apparatus, method, and computer program product are disclosed. The apparatus includes a processor and a memory that stores code executable by the processor, including code that accesses personal information of a user, identifies a dialectal nuance of the user based on the personal information, and selects a word recognition dictionary based on the dialectal nuance. The dialectal nuance may be based on a location of the user, a nationality of the user, an age of the user, an education level of the user, and/or a profession of the user. The apparatus may also suggest one or more text entries from the selected word recognition dictionary based on the user input.

User Modeling for a Personal Assistant

Abstract: We present a user modeling system that serves as the foundation of a personal assistant. The system ingests web search history for signed-in users, and identifies coherent contexts that correspond to tasks, interests, and habits. Unlike past work which focused on either in-session tasks or tasks over a few days, we look at several months of history in order to identify not just short-term tasks, but also long-term interests and habits. The features we use for identifying coherent contexts yield substantially higher precision and recall than past work. We also present an algorithm for identifying contexts that is 8 to 30 times faster than previous algorithms. The user modeling system has been deployed in production. It runs over hundreds of millions of users, and updates the models with a 10-minute latency. The contexts identified by the system serve as the foundation for generating recommendations in Google Now.

User-Specific Feature-Based Similarity Models for Top-n Recommendation of New Items

Abstract: Recommending new items for suitable users is an important yet challenging problem due to the lack of preference history for the new items. Noncollaborative user modeling techniques that rely on the item features can be used to recommend new items. However, they only use the past preferences of each user to provide recommendations for that user. They do not utilize information from the past preferences of other users, which can potentially be ignoring useful information. More recent factor models transfer knowledge across users using their preference information in order to provide more accurate recommendations. These methods learn a low-rank approximation for the preference matrix, which can lead to loss of information. Moreover, they might not be able to learn useful patterns given very sparse datasets. In this work, we present UFSM, a method for top-n recommendation of new items given binary user preferences. UFSM learns User-specific Feature-based item-Similarity Models, and its strength lies in combining two points: (1) exploiting preference information across all users to learn multiple global item similarity functions and (2) learning user-specific weights that determine the contribution of each global similarity function in generating recommendations for each user. UFSM can be considered as a sparse high-dimensional factor model where the previous preferences of each user are incorporated within his or her latent representation. This way, UFSM combines the merits of item similarity models that capture local relations among items and factor models that learn global preference patterns. A comprehensive set of experiments was conduced to compare UFSM against state-of-the-art collaborative factor models and noncollaborative user modeling techniques. Results show that UFSM outperforms other techniques in terms of recommendation quality. UFSM manages to yield better recommendations even with very sparse datasets. Results also show that UFSM can efficiently handle high-dimensional as well as low-dimensional item feature spaces.

Device voice control

Abstract: Systems and processes for device voice control are provided. An example process includes, at an electronic device, receiving a spoken user input and interpreting the spoken user input to derive a representation of user intent. The process further includes determining whether a task may be identified based on the representation of user intent. In accordance with a determination that a task may be identified based on the representation of user intent, the task is performed, and in accordance with a determination that a task may not be identified based on the representation of user intent, the spoken user input is disambiguated.

Fuzzy Logic for Adaptive Instruction in an E-learning Environment for Computer Programming

Abstract: In this paper, a novel approach to web-based education that performs individualized instruction on the domain of programming languages is presented. This approach is fully implemented and evaluated in an educational application module, called fuzzy knowledge state definer (FuzKSD). In particular, FuzKSD performs user modeling by dynamically identifying and updating a student's knowledge level of all the concepts of the domain knowledge. The operation of FuzKSD is based on fuzzy cognitive maps (FCMs) that are used to represent the dependences among the domain concepts. FuzKSD uses fuzzy sets to represent a student's knowledge level as a subset of the domain knowledge. Thus, it combines fuzzy theory with the overlay model. Moreover, it employs a novel inference mechanism that dynamically updates user stereotypes using fuzzy sets. It should be noted that the overlay model and stereotypes constitute two widely used methods for user modeling. However, they have not been combined with fuzzy sets thus far in the literature. The gain from this novel combination is significant as a student level of knowledge is represented in a more realistic way by automatically modeling the learning or forgetting process of a student with respect to the FCMs and thus the system can provide individualized adaptive advice. The application of this approach is not limited to adaptive instruction. It can also be used in other systems with changeable user states, such as e-shops, where consumers' preferences change over time and affect one another. Therefore, the particular module constitutes a novel generic fuzzy tool, which offers dynamic adaptation to users' needs and preferences of adaptive systems.

When-To-Post on Social Networks

Abstract: For many users on social networks, one of the goals when broadcasting content is to reach a large audience. The probability of receiving reactions to a message differs for each user and depends on various factors, such as location, daily and weekly behavior patterns and the visibility of the message. While previous work has focused on overall network dynamics and message flow cascades, the problem of recommending personalized posting times has remained an under-explored topic of research.In this study, we formulate a when-to-post problem, where the objective is to find the best times for a user to post on social networks in order to maximize the probability of audience responses. To understand the complexity of the problem, we examine user behavior in terms of post-to-reaction times, and compare cross-network and cross-city weekly reaction behavior for users in different cities, on both Twitter and Facebook. We perform this analysis on over a billion posted messages and observed reactions, and propose multiple approaches for generating personalized posting schedules. We empirically assess these schedules on a sampled user set of 0.5 million active users and more than 25 million messages observed over a 56 day period. We show that users see a reaction gain of up to 17% on Facebook and 4% on Twitter when the recommended posting times are used.We open the dataset used in this study, which includes timestamps for over 144 million posts and over 1.1 billion reactions. The personalized schedules derived here are used in a fully deployed production system to recommend posting times for millions of users every day.

Modelling the persuadee in asymmetric argumentation dialogues for persuasion

Abstract: Computational models of argument could play a valuable role in persuasion technologies for behaviour change (e.g. persuading a user to eat a more healthy diet, or to drink less, or to take more exercise, or to study more conscientiously, etc). For this, the system (the persuader) could present arguments to convince the user (the persuadee). In this paper, we consider asymmetric dialogues where only the system presents arguments, and the system maintains a model of the user to determine the best choice of arguments to present (including counterarguments to key arguments believed to be held by the user). The focus of the paper is on the user model, including how we update it as the dialogue progresses, and how we use it to make optimal choices for dialogue moves.

User Profiling Trends, Techniques and Applications

Abstract: The Personalization of information has taken recommender systems at a very high level. With personalization these systems can generate user specific recommendations accurately and efficiently. User profiling helps personalization, where information retrieval is done to personalize a scenario which maintains a separate user profile for individual user. The main objective of this paper is to explore this field of personalization in context of user profiling, to help researchers make aware of the user profiling. Various trends, techniques and Applications have been discussed in paper which will fulfill this motto.

Adapting Recommendations to Contextual Changes Using Hierarchical Hidden Markov Models

Abstract: Recommender systems help users find items of interest by tailoring their recommendations to users' personal preferences. The utility of an item for a user, however, may vary greatly depending on that user's specific situation or the context in which the item is used. Without considering these changes in preferences, the recommendations may match the general preferences of a user, but they may have small value for the user in his/her current situation. In this paper, we introduce a hierarchical hidden Markov model for capturing changes in user's preferences. Using a user's feedback sequence on items, we model the user as a hierarchical hidden Markov process and the current context of the user as a hidden variable in this model. For a given user, our model is used to infer the maximum likelihood sequence of transitions between contextual states and to predict the probability distribution for the context of the next action. The predicted context is then used to generate recommendations. Our evaluation results using Last.fm music playlist data, indicate that this approach achieves significantly better performance in terms of accuracy and diversity compared to baseline methods.

Evaluation of topic-based adaptation and student modeling in QuizGuide

Abstract: This paper presents an in-depth analysis of a nonconventional topic-based personalization approach for adaptive educational systems (AES) that we have explored for a number of years in the context of university programming courses. With this approach both student modeling and adaptation are based on coarse-grained knowledge units that we called topics. Our motivation for the topic-based personalization was to enhance AES transparency for both teachers and students by utilizing typical topic-based course structures as the foundation for designing all aspects of an AES from the domain model to the end-user interface. We illustrate the details of the topic-based personalization technology, with the help of the Web-based educational service QuizGuide--the first system to implement it. QuizGuide applies the topic-based personalization to guide students to the right learning material in the context of an undergraduate C programming course. While having a number of architectural and practical advantages, the suggested coarse-grained personalization approach deviates from the common practices toward knowledge modeling in AES. Therefore, we believe that several aspects of QuizGuide required a detailed evaluation--from modeling accuracy to the effectiveness of adaptation. The paper discusses how this new student modeling approach can be evaluated, and presents our attempts to evaluate it from multiple different prospects. The evaluation of QuizGuide across several consecutive semesters demonstrates that, although topics do not always support precise user modeling, they can provide a basis for successful personalization in AESs.

Dynamic user profiles for web personalisation

Abstract: We describe two algorithms for improving the mapping of interests to an ontology.We develop methods for modelling short and long-term user profiles.We introduce methods for adapting user profiles based on ongoing user behaviour.We demonstrate the effectiveness of our approach in a personalised search system. Web personalisation systems are used to enhance the user experience by providing tailor-made services based on the user's interests and preferences which are typically stored in user profiles. For such systems to remain effective, the profiles need to be able to adapt and reflect the users' changing behaviour. In this paper, we introduce a set of methods designed to capture and track user interests and maintain dynamic user profiles within a personalisation system. User interests are represented as ontological concepts which are constructed by mapping web pages visited by a user to a reference ontology and are subsequently used to learn short-term and long-term interests. A multi-agent system facilitates and coordinates the capture, storage, management and adaptation of user interests. We propose a search system that utilises our dynamic user profile to provide a personalised search experience. We present a series of experiments that show how our system can effectively model a dynamic user profile and is capable of learning and adapting to different user browsing behaviours.

Towards a comprehensive framework for QOE and user behavior modelling

Abstract: While the modeling of QoE has made significant advances over the last couple of years, currently existing models still lack an integration of user behavior aspects and user context factors along with the consideration of appropriate temporal scales Therefore, the goal of this paper is to present a comprehensive QoE and user behavior model providing a framework which allows joining a multitude of existing modeling approaches under the perspectives of service provider benefit, user well-being and technical system performance In addition, we discuss the role of a broad range of corresponding influence factors, with a specific emphasis on user and context issues, and illustrate our proposal through a series of related use cases

Tailoring Music Recommendations to Users by Considering Diversity, Mainstreaminess, and Novelty

Abstract: A shortcoming of current approaches for music recommendation is that they consider user-specific characteristics only on a very simple level, typically as some kind of interaction between users and items when employing collaborative filtering. To alleviate this issue, we propose several user features that model aspects of the user's music listening behavior: diversity, mainstreaminess, and novelty of the user's music taste. To validate the proposed features, we conduct a comprehensive evaluation of a variety of music recommendation approaches (stand-alone and hybrids) on a collection of almost 200 million listening events gathered from \propername{Last.fm}. We report first results and highlight cases where our diversity, mainstreaminess, and novelty features can be beneficially integrated into music recommender systems.

Getting More for Less: Optimized Crowdsourcing with Dynamic Tasks and Goals

Abstract: In crowdsourcing systems, the interests of contributing participants and system stakeholders are often not fully aligned. Participants seek to learn, be entertained, and perform easy tasks, which offer them instant gratification; system stakeholders want users to complete more difficult tasks, which bring higher value to the crowdsourced application. We directly address this problem by presenting techniques that optimize the crowdsourcing process by jointly maximizing the user longevity in the system and the true value that the system derives from user participation. We first present models that predict the "survival probability" of a user at any given moment, that is, the probability that a user will proceed to the next task offered by the system. We then leverage this survival model to dynamically decide what task to assign and what motivating goals to present to the user. This allows us to jointly optimize for the short term (getting difficult tasks done) and for the long term (keeping users engaged for longer periods of time). We show that dynamically assigning tasks significantly increases the value of a crowdsourcing system. In an extensive empirical evaluation, we observed that our task allocation strategy increases the amount of information collected by up to 117.8%. We also explore the utility of motivating users with goals. We demonstrate that setting specific, static goals can be highly detrimental to the long-term user participation, as the completion of a goal (e.g., earning a badge) is also a common drop-off point for many users. We show that setting the goals dynamically, in conjunction with judicious allocation of tasks, increases the amount of information collected by the crowdsourcing system by up to 249%, compared to the existing baselines that use fixed objectives.

Computational Model for Mood

Abstract: According to some aspects, disclosed methods and systems may include having a user input one or more speech commands into an input device of a user device. The user device may communicate with one or more components or devices at a local office or headend. The local office or the user device may transcribe the speech commands into language transcriptions. The local office or the user device may determine a mood for the user based on whether any of the speech commands may have been repeated. The local office or the user device may determine, based on the mood of the user, which content asset or content service to make available to the user device.