Abstract: In offering personalized content geared toward users’ individual interests, recommender systems are assumed to reduce news diversity and thus lead to partial information blindness (i.e., filter bubbles). We conducted two exploratory studies to test the effect of both implicit and explicit personalization on the content and source diversity of Google News. Except for small effects of implicit personalization on content diversity, we found no support for the filter-bubble hypothesis. We did, however, find a general bias in that Google News over-represents certain news outlets and under-represents other, highly frequented, news outlets. The results add to a growing body of evidence, which suggests that concerns about algorithmic filter bubbles in the context of online news might be exaggerated.

Abstract: There is a growing consensus that online platforms have a systematic influence on the democratic process. However, research beyond social media is limited. In this paper, we report the results of a mixed-methods algorithm audit of partisan audience bias and personalization within Google Search. Following Donald Trump's inauguration, we recruited 187 participants to complete a survey and install a browser extension that enabled us to collect Search Engine Results Pages (SERPs) from their computers. To quantify partisan audience bias, we developed a domain-level score by leveraging the sharing propensities of registered voters on a large Twitter panel. We found little evidence for the "filter bubble'' hypothesis. Instead, we found that results positioned toward the bottom of Google SERPs were more left-leaning than results positioned toward the top, and that the direction and magnitude of overall lean varied by search query, component type (e.g. "answer boxes"), and other factors. Utilizing rank-weighted metrics that we adapted from prior work, we also found that Google's rankings shifted the average lean of SERPs to the right of their unweighted average.

Abstract: The purpose of this paper is to examine the impact of perceived social media marketing activities (SMMAs) on customer loyalty via customer equity drivers (CEDs) in an e-commerce context.,The study surveyed 371 students from a large university in India. The data were analyzed via confirmatory factor analysis and the research hypotheses were examined using SEM.,The study revealed three key findings. First, perceived SMMAs of e-commerce comprise five dimensions, namely, interactivity, informativeness, word-of-mouth, personalization and trendiness. Second, perceived SMMAs of e-commerce have significantly and positively influenced all the drivers of customer equity (CEDs). Third, the CEDs of e-commerce exhibit a significant and positive influence on customer loyalty toward the e-commerce sites.,This study will help e-commerce managers to boost customer loyalty toward the e-commerce sites through perceived SMMA.,The study is the first to identify five dimensions of e-commerce perceived SMMA. The current study also introduces the stimulus–organism–response model as a theoretical support to connect perceived SMMAs of e-commerce to customers’ loyalty via CEDs. This is supposed to be the first study to examine the impact of perceived SMMA on customer loyalty toward the e-commerce sites via CEDs in the e-commerce industry.

Abstract: In collaboration with three companies selling a diverse set of products, we conducted randomized field experiments in which experimentally tailored email ads were sent to millions of individuals. We found consistently that personalizing the emails by adding consumer-specific information e.g., recipient's name benefited the advertisers. Importantly, such content is not likely to be informative about the advertised product or the company. In our main experiment, we found that adding the name of the message recipient to the email's subject line increased the probability of the recipient opening it by 20% from 9.05% to 10.80%, which translated to an increase in sales' leads by 31% from 0.39% to 0.51% and a reduction in the number of individuals unsubscribing from the email campaign by 17% from 1.2% to 1.0%. We present similar experiments conducted with other companies, which show that the effects we document extend from objectives ranging from acquiring new customers to retaining customers who have purchased from the company in the past. Our investigation of several possible mechanisms suggests that such content increases the effort consumers make in processing the other content in the rest of the advertising message. Our paper quantifies the benefits from personalization and sheds light on the role of noninformative advertising content by analyzing several detailed measures of recipient's interaction with the message. It provides external validity to psychological mechanisms and has clear implications for the firms that are designing their advertising campaigns. Data and the online appendix are available at https://doi.org/10.1287/mksc.2017.1066 .

Abstract: Search engines are a primary means through which people obtain information in today»s connected world. Yet, apart from the search engine companies themselves, little is known about how their algorithms filter, rank, and present the web to users. This question is especially pertinent with respect to political queries, given growing concerns about filter bubbles, and the recent finding that bias or favoritism in search rankings can influence voting behavior. In this study, we conduct a targeted algorithm audit of Google Search using a dynamic set of political queries. We designed a Chrome extension to survey participants and collect the Search Engine Results Pages (SERPs) and autocomplete suggestions that they would have been exposed to while searching our set of political queries during the month after Donald Trump»s Presidential inauguration. Using this data, we found significant differences in the composition and personalization of politically-related SERPs by query type, subjects» characteristics, and date.

Abstract: The privacy calculus suggests that online self-disclosure is based on a cost–benefit trade-off. However, although companies progressively collect information to offer tailored services, the effect of both personalization and context-dependency on self-disclosure has remained understudied. Building on the privacy calculus, we hypothesized that benefits, privacy costs, and trust would predict online self-disclosure. Moreover, we analyzed the impact of personalization, investigating whether effects would differ for health, news, and commercial websites. Results from an online experiment using a representative Dutch sample (N = 1,131) supported the privacy calculus, revealing that it was stable across contexts. Personalization decreased trust slightly and benefits marginally. Interestingly, these effects were context-dependent: While personalization affected outcomes in news and commerce contexts, no effects emerged in the health context.

Abstract: In e-learning, most content-based (CB) recommender systems provide recommendations depending on matching rules between learners and learning objects (LOs). Such learner-oriented approaches are limited when it comes to detecting learners’ changes, furthermore, the recommendations show low adaptability and diversity. In this study, in order to improve the adaptability and diversity of recommendations, we incorporate an LO-oriented recommendation mechanism to learner-oriented recommender systems, and propose an LO self-organization based recommendation approach (Self). LO self-organization means LO interacts with each other in a spontaneous and autonomous way. Such self-organization behavior is conducive to generating a stable LO structure through information propagation. The proposed approach works as follows: firstly, LOs are simulated as intelligent entities using the self-organization theory. LOs can receive information, transmit information, as well as move. Secondly, an environment perception module is designed. This module can capture and perceive learner’s preference drifts by analyzing LOs’ self-organization behaviors. Finally, according to learners’ explicit requirements and implicit preference drifts, recommendations are generated through LOs’ self-organization behaviors. Based on applications to real-life learning processes, the ample experimental results demonstrate the high adaptability, diversity, and personalization of the recommendations.

Abstract: Self-reflection is a central goal of personal informatics systems, and constructing visualizations from physical tokens has been found to help people reflect on data. However, so far, constructive physicalization has only been studied in lab environments with provided datasets. Our qualitative study investigates the construction of personal physicalizations in people's domestic environments over 2-4 weeks. It contributes an understanding of (1) the process of creating personal physicalizations, (2) the types of personal insights facilitated, (3) the integration of self-reflection in the physicalization process, and (4) its benefits and challenges for self-reflection. We found that in constructive personal physicalization, data collection, construction and self-reflections are deeply intertwined. This extends previous models of visualization creation and data-driven self-reflection. We outline how benefits such as reflection through manual construction, personalization, and presence in everyday life can be transferred to a wider set of digital and physical systems.

Abstract: Personalization of the e-learning systems according to the learner’s needs and knowledge level presents the key element in a learning process. E-learning systems with personalized recommendations should adapt the learning experience according to the goals of the individual learner. Aiming to facilitate personalization of a learning content, various kinds of techniques can be applied. Collaborative and social tagging techniques could be useful for enhancing recommendation of learning resources. In this paper, we analyze the suitability of different techniques for applying tag-based recommendations in e-learning environments. The most appropriate model ranking, based on tensor factorization technique, has been modified to gain the most efficient recommendation results. We propose reducing tag space with clustering technique based on learning style model, in order to improve execution time and decrease memory requirements, while preserving the quality of the recommendations. Such reduced model for providing tag-based recommendations has been used and evaluated in a programming tutoring system.

Abstract: Personalization is one of the recurring themes in education and has occupied a large amount of specialised literature, since its appearance in the 1960s. A systematic exploration of the literature of the last 55 years (1960–2015) is presented and is intended to analyse which educational perspective underlies the customized environments or experiences proposed in the educational technology that is addressed in the literature. It is important to understand that this analysis is a very relevant challenge, if we want to understand what pedagogical approaches have been continuously developed and how and why we should consider their future. The results show a complete centralisation of experiences in technological developments, the majority of them focussed in Higher Education, as well as a lack of an explicit pedagogical perspective in the experiences analysed, especially those with greater impact. It also shows a shortage of in-house pedagogical material – developed in the light of this research, that evolves and makes an impact on the educational landscape.

Abstract: For many years, the main goal of the Netflix personalized recommendation system has been to get the right titles in front of our members at the right time. But the job of recommendation does not end there. The homepage should be able to convey to the member enough evidence of why a title may be good for her, especially for shows that the member has never heard of. One way to address this challenge is to personalize the way we portray the titles on our service. An important aspect of how to portray titles is through the artwork or imagery we display to visually represent each title. The artwork may highlight an actor that you recognize, capture an exciting moment like a car chase, or contain a dramatic scene that conveys the essence of a movie or show. It is important to select good artwork because it may be the first time a member becomes aware of a title (and sometimes the only time), so it must speak to them in a meaningful way. In this talk, we will present an approach for personalizing the artwork we use on the Netflix homepage. The system selects an image for each member and video to give better visual evidence for why the title might be appealing to that particular member. There are many challenges involved in getting artwork personalization to succeed. One challenge is that we can only select a single piece of artwork to represent each title. In contrast, typical recommendation engines present multiple items (in some order) to a member allowing us to subsequently learn about preferences between items through the specific item a member selects from the presented assortment. In contrast, we only collect feedback from the one image that was presented to each member for each title. This leads to a training paradigm based on incomplete logged bandit feedback [1]. Moreover, since the artwork selection process happens on top of a recommendation system, collecting data directly from the production experience (observational data) makes it hard to detangle whether a play was due to the recommendation or from the incremental effect of personalized evidence. Another challenge is understanding the impact of changing the artwork between sessions and if that is beneficial or confusing to the user. We also need to consider how diverse artworks perform in relation to one another. Finally, given that the popularity and audiences for titles can change or drop quickly after launch, the system needs to quickly learn how to personalize images for a new item. All these considerations naturally lead us to frame the problem as online learning with contextual multi-arm bandits. Briefly, contextual bandits are a class of online learning algorithms that balance the cost of gathering randomized training data (which is required for learning an unbiased model on an ongoing basis) with the benefits of applying the learned model to each member context (to maximize user engagement). This is known as the explore-exploit trade-off. In this setting, for a given title the set of actions is the set of available images for the title. We aim to discover the underlying unknown reward, based on probability of play, for each image given a member, a title, and some context. The context could be based on profile attributes (geo-localization, previous plays, etc), the device, time, and other factors that might affect what is the optimal image to choose in each session. With a large member base, many titles in the catalog, and multiple images per title, Netflix's product is an ideal platform to test ideas for personalization of artwork. At peak, over 20 million personalized image requests per second need to be handled with low latency. To train our model, we leveraged existing logged data from a previous system that chose images in an unpersonalized manner. We will present results comparing the contextual bandit personalization algorithms using offline policy evaluation metrics [2], such as inverse propensity scoring and doubly robust estimators [3]. We will conclude with a discussion of opportunities to expand and improve our approach. This includes developing algorithms to handle cold-start by quickly personalizing new images and new titles. We also discuss extending this personalization approach across other types of artwork we use and other evidence that describe our titles such as synopses, metadata, and trailers. Finally, we discuss potentially closing the loop by looking at how we can help artists and designers figure out what new imagery they should create to make a title even more compelling and personalizable.

Abstract: We consider an online retailer facing heterogeneous customers with initially unknown product preferences. Customers are characterized by a diverse set of demographic and transactional attributes. T...

Abstract: Currently, food image recognition tasks are evaluated against fixed datasets. However, in real-world conditions, there are cases in which the number of samples in each class continues to increase and samples from novel classes appear. In particular, dynamic datasets in which each individual user creates samples and continues the updating process often has content that varies considerably between different users, and the number of samples per person is very limited. A single classifier common to all users cannot handle such dynamic data. Bridging the gap between the laboratory environment and the real world has not yet been accomplished on a large scale. Personalizing a classifier incrementally for each user is a promising way to do this. In this paper, we address the personalization problem, which involves adapting to the user's domain incrementally using a very limited number of samples. We propose a simple yet effective personalization framework, which is a combination of the nearest class mean classifier and the 1-nearest neighbor classifier based on deep features. To conduct realistic experiments, we made use of a new dataset of daily food images collected by a food-logging application. Experimental results show that our proposed method significantly outperforms existing methods.

Abstract: Mobile users spend an extensive amount of time on typing. A more efficient text input instrument brings a significant enhancement of user experience. Deep learning techniques have been recently applied to suggesting the next words of input, but to achieve more accurate predictions, these models should be customized for individual users. Personalization is often at the expense of privacy concerns. Existing solutions require users to upload the historical logs of their input text to the cloud so that a deep learning predictor can be trained. In this work, we propose a novel approach, called DeepType, to personalize text input with better privacy. The basic idea is intuitive: training deep learning predictors on the device instead of on the cloud, so that the model makes personalized and private data never leaves the device to externals. With DeepType, a global model is first trained on the cloud using massive public corpora, and our personalization is done by incrementally customizing the global model with data on individual devices. We further propose a set of techniques that effectively reduce the computation cost of training deep learning models on mobile devices at the cost of negligible accuracy loss. Experiments using real-world text input from millions of users demonstrate that DeepType significantly improves the input efficiency for individual users, and its incurred computation and energy costs are within the performance and battery restrictions of typical COTS mobile devices.

Abstract: The complexity of today’s learning processes and practices entails various challenges. It is becoming much harder for teachers to observe, control, and adjust the learning process. Moreover, contemporary teaching is enhanced with different technologies and systems that not only support information-transfer, but also make this process more effective. In this paper we present the Programming Tutoring System (ProTuS), which provides smart and interactive content, personalization options, adaptive features, and learning analytics as a support for users engaged in learning complex cognitive skills. Our contribution in this paper is twofold, conceptual and empirical. The paper presents the interactive learning analytics component developed in ProTuS and the results from the empirical study. The study shows that students find adaptive learning systems to be useful in monitoring progress, promoting reflective practices, and receiving feedback to better understand their actions and learning strategies.

Abstract: With the increasing popularity of smartphones and tablets, game-based learning (GBL) is undergoing a rapid shift to mobile platforms. This transformation is driven by mobility, wireless interfaces, and built-in sensors that these smart devices offer in order to enable blended and context-sensitive mobile learning (m-Learning) activities. Thus, m-Learning is becoming more independent and ubiquitous (u-Learning). In order to identify and analyze the main trends and the future challenging issues involved in designing mGBL learning strategies, as well as to bring to the foreground important issues pertaining to mobile and context-aware ubiquitous GBL, the work at hand conducts a comprehensive survey of this particular area. Specifically, it introduces and applies a six-dimensional framework consisted of Spatio-temporal, Collaboration/Social, Session, Personalization, Data security & privacy, and Pedagogy, with the aim of scrutinizing the contributions in the field of mGBL published from 2004 to 2016. It was found that the transition to mGBL presents several difficulties, and therefore cannot be conceived as a simple and quick modification of existing GBL solutions. In this respect, this work is anticipated to foster the development of well-designed solutions that are intensive not only in their technological aspect, but in pedagogical qualities as well.

Abstract: Knowledge graphs have been used to support a wide range of applications and enhance search results for multiple major search engines, such as Google and Bing. At Amazon we are building a Product Graph, an authoritative knowledge graph for all products in the world. The thousands of product verticals we need to model, the vast number of data sources we need to extract knowledge from, the huge volume of new products we need to handle every day, and the various applications in Search, Discovery, Personalization, Voice, that we wish to support, all present big challenges in constructing such a graph. In this talk we describe four scientific directions we are investigating in building and using such a knowledge graph. First, we have been developing advanced extraction technologies to harvest product knowledge from semi-structured sources on the web and from text product profiles. Our annotation-based extraction tool selects a few webpages (typically below 10 pages) from a website for annotations, and can derive XPaths to extract from the whole website with average precision and recall of 97% [1]. Our distantly supervised extraction tool, CERES, uses an existing knowledge graph to automatically generate (noisy) training labels, and can obtain a precision over 90% when extracting from long-tail websites in various languages [1]. Our OpenTag technique extends state-of-the-art techniques such as Recursive Neural Network (RNN) and Conditional Random Field with attention and active learning, to achieve over 90% precision and recall in extracting attribute values (including values unseen in training data) from product titles, descriptions, and bullets [3].

Abstract: To assist consumers in product search and selection while shopping online, many e-commerce retailers have implemented web-based product recommendation agents (PRAs). However, consumers are empowered to the extent that the PRAs provide true personalization by recommending products based solely on, and thus best representing, consumers' preferences. This study constructs and empirically tests a theoretical model that examines how biased recommendations from PRAs influence consumers' decision quality and decision effort. The results of an online experiment show that consumers are extremely vulnerable to biased personalized recommendations from online PRAs. In addition, our results extend prior research by identifying perceived personalization as a critical mechanism driving the influence of biased PRA on consumers' decision quality and decision effort. This study fills a void in the literature and calls attention to an insidious form of manipulation made possible by innovative technologies supporting e-commerce.

Abstract: In many settings, a decision-maker wishes to learn a rule, or policy, that maps from observable characteristics of an individual to an action. Examples include selecting offers, prices, advertisements, or emails to send to consumers, as well as the problem of determining which medication to prescribe to a patient. While there is a growing body of literature devoted to this problem, most existing results are focused on the case where data comes from a randomized experiment, and further, there are only two possible actions, such as giving a drug to a patient or not. In this paper, we study the offline multi-action policy learning problem with observational data and where the policy may need to respect budget constraints or belong to a restricted policy class such as decision trees. We build on the theory of efficient semi-parametric inference in order to propose and implement a policy learning algorithm that achieves asymptotically minimax-optimal regret. To the best of our knowledge, this is the first result of this type in the multi-action setup, and it provides a substantial performance improvement over the existing learning algorithms. We then consider additional computational challenges that arise in implementing our method for the case where the policy is restricted to take the form of a decision tree. We propose two different approaches, one using a mixed integer program formulation and the other using a tree-search based algorithm.

Abstract: Recommender systems have shown to be valuable tools for filtering, ranking, and discovery in a variety of application domains such as e-commerce, media repositories or document-based information in general that includes the various scenarios of Social Information Access discussed in this book. One key to the success of such systems lies in the precise acquisition or estimation of the user’s preferences. While general recommender systems research often relies on the existence of explicit preference statements for personalization, such information is often very sparse or unavailable in real-world applications. Information that allows us to assess the relevance of certain items indirectly through a user’s actions and behavior (implicit feedback) is in contrast often available in abundance. In this chapter we categorize different types of implicit feedback and review their use in the context of recommender systems and Social Information Access applications. We then extend the categorization scheme to be suitable to recent application domains. Finally, we present state-of-the-art algorithmic approaches, discuss challenges when using implicit feedback signals in particular with respect to popularity biases, and discuss selected recent works from the literature.

Abstract: Fueled by advancing technologies that continually expand Web data tracking and aggregating capabilities, online advertising has become increasingly personalized and pervasive. This trend is largely responsible for a growing number of consumers (more than 615 million worldwide) choosing to install ad-blocking software on their computers and mobile devices. As a result, U.S. publishers and advertisers estimate that ad blockers cost them more than $28 billion in revenue in the first half of 2017, and this figure is forecast to exceed $35 billion by 2020. Rooted in a theoretical foundation of psychological reactance theory (PRT), the present study examines key cognitive and affective factors driving consumers to reject personalized advertising messages and install ad-blocking software. A structural equation analysis reveals privacy-related threats, along with benefits rooted in relevance and rewards, moderated by the type of data being used to personalize advertising messages, are contributing to this...

Abstract: The pervasive availability of voice assistants may support children in finding answers to informational queries by removing the literacy requirements of text search (e.g., typing, spelling). However, most such systems are not designed for the specific needs and preferences of children and may struggle with understanding the intent of their questions. In our investigation, we observed 87 children and 27 adults interacting with three Wizard-of-Oz speech interfaces to arrive at answers to questions that required reformulation. We found that many children and some adults required help to reach an effective question reformulation. We report the common types of reformulations (both effective and ineffective ones). We also compared three versions of speech interfaces with different approaches to referring to itself (personification) and to the participant (naming personalization). We found that children preferred personified interfaces, but naming personalization did not affect preference. We connect our findings to implications for design of speech systems for families.

Abstract: This chapter examines how tailored marketing strategies in the form of customization and personalization can foster customer engagement along the customer lifecycle. Building on various real-world examples, the authors outline the general array of opportunities that companies have at their disposal to customize and/or personalize the marketing mix. In addition, they discuss important contingency factors that facilitate or impede the applicability of both strategies under various business circumstances.

Abstract: The act of reading has benefits for individuals and societies, yet studies show that reading declines, especially among the young. Recommender systems can help stop such decline. We present a survey of recommender systems in the domain of books. We have categorized the systems into six classes, and highlighted the main trends, issues, evaluation approaches and datasets. Other research areas, such as psychology, are consulted to understand users' books choices and reading models.

Abstract: A multitude of challenges comes into play when attempting to design (and evaluate) an interactive digital storytelling experience for use by visitors in a museum. This paper reports on the evaluation of the prototype mobile-based storytelling “guides” designed, developed and deployed as part of a research project at the Acropolis Museum in Athens, Greece. Experiences designed for different visitor profiles were evaluated several times throughout the iterative design process, in a number of on-site studies, and with more than 180 museum visitors of all ages (with this paper reporting on two studies conducted with a total of 53 users visiting individually or in pairs). The evaluation methods included ethnography (i.e., observation of visitors in the Museum’s galleries), pre- and post-experience in-depth interviews and questionnaires to measure the Users’ Experience (UX), as well as data logging. The analysis of the data focused on themes representing components of the experiences, such as interactive story plot and narration, staging and way-finding in the physical space, personalization and social interaction. Our findings confirmed that understanding UX and what makes it effective or not in the rich context of a cultural setting is a complex endeavor. The paper discusses our findings and proposes relevant recommendations for the design of digital experiences for cultural, educational, and recreational purposes.

Abstract: Collaborative Filtering Recommender Systems predict user preferences for online information, products or services by learning from past user-item relationships. A predominant approach to Collaborative Filtering is Neighborhood-based, where a user-item preference rating is computed from ratings of similar items and/or users. This approach encounters data sparsity and scalability limitations as the volume of accessible information and the active users continue to grow leading to performance degradation, poor quality recommendations and inaccurate predictions. Despite these drawbacks, the problem of information overload has led to great interests in personalization techniques. The incorporation of context information and Matrix and Tensor Factorization techniques have proved to be a promising solution to some of these challenges. We conducted a focused review of literature in the areas of Context-aware Recommender Systems utilizing Matrix Factorization approaches. This survey paper presents a detailed literature review of Context-aware Recommender Systems and approaches to improving performance for large scale datasets and the impact of incorporating contextual information on the quality and accuracy of the recommendation. The results of this survey can be used as a basic reference for improving and optimizing existing Context-aware Collaborative Filtering based Recommender Systems. The main contribution of this paper is a survey of Matrix Factorization techniques for Context-aware Collaborative Filtering Recommender Systems.

Abstract: Purpose Personalized medicine is expected to yield improved health outcomes. Data mining over massive volumes of patients' clinical data is an appealing, low-cost and noninvasive approach toward personalization. Machine learning algorithms could be trained over clinical "big data" to build prediction models for personalized therapy. To reach this goal, a scalable "big data" architecture for the medical domain becomes essential, based on data standardization to transform clinical data into FAIR (Findable, Accessible, Interoperable and Reusable) data. Using Ontologies and Semantic Web technologies, we attempt to reach mentioned goal. Methods We developed an ontology to be used in the field of radiation oncology to map clinical data from relational databases. We combined ontology with semantic Web techniques to publish mapped data and easily query them using SPARQL. Results The Radiation Oncology Ontology (ROO) contains 1,183 classes and 211 properties between classes to represent clinical data (and their relationships) in the radiation oncology domain following FAIR principles. We combined the ontology with Semantic Web technologies showing how to efficiently and easily integrate and query data from different (relational database) sources without a priori knowledge of their structures. Discussion When clinical FAIR data sources are combined (linked data) using mentioned technologies, new relationships between entities are created and discovered, representing a dynamic body of knowledge that is continuously accessible and increasing.