HCI Bibliography Home | HCI Conferences | About RecSys | RecSys Conf Proceedings | Detailed Records | RefWorks | EndNote | Hide Abstracts
RecSys Tables of Contents: 0708091011121314 ⇐ MORE

Proceedings of the 2012 ACM Conference on Recommender Systems

Fullname:Proceedings of the Sixth ACM Conference on Recommender Systems
Editors:Pádraig Cunningham; Neil Hurley; Ido Guy; Sarabjot Singh Anand
Location:Dublin, Ireland
Dates:2012-Sep-09 to 2012-Sep-13
Publisher:ACM
Standard No:ISBN: 978-1-4503-1270-7; ACM DL: Table of Contents; hcibib: RecSys12
Papers:79
Pages:360
Links:Conference Website
Summary:It is our great pleasure to welcome you to the 6th ACM Recommender Systems Conference (ACM RecSys 2012), held in Dublin, Ireland on September 9-13. As RecSys enters the second half of its first decade, it has clearly established itself as the premier international venue for research and development in the field of Recommender Systems, where leading researchers and practitioners from around the world meet and discuss their latest results and solutions.
    The growing maturity of the conference is reflected through several changes in the program as compared to previous years: the tutorial program includes four exciting tutorials, which, for the first time, were not invited, but rather submitted and rigorously reviewed. We also put a special emphasis on soliciting demonstrations, leading to a record number of 10 demos accepted to the conference. In addition, this year continues the tradition of attractive collocated workshops, with a total of 8. Finally, 8 PhD students have been accepted to participate in the Doctoral Symposium. In parallel to growing and diversifying participation opportunities, the selectivity of the short paper track was substantially increased, while the competitive acceptance rate of the long paper track was maintained.
    Overall, the number of paper submissions went up almost 10% as compared to last year, with a total of 185 submissions. Out of 119 full paper submissions, 24 were accepted (20.2%) for oral presentation at the conference. Out of 66 short paper submissions, 21 were accepted (31.8%) for poster presentation at the conference. 5 of these 21 have been identified as industry short papers, highlighting the fact that their main contribution lies in the description of a real system, typically already in wide use.
    The conference program includes two keynotes, one from an academic perspective by Jure Leskovec (Stanford University) and one from an industrial perspective by Ron Kohavi (Microsoft). The Industry program also includes a rich set of talks by Ralf Herbrich (Facebook), Ronny Lempel (Yahoo! Research), Sumanth Kolar (StumbleUpon), Anmol Bhasin (LinkedIn), Thore Graepel (Microsoft Research), and Paul Lamere (The Echo Nest).
  1. Keynote address
  2. Tutorials
  3. Multi-objective recommendation and human factors
  4. Social recommendation
  5. Implicit feedback and user preference
  6. Contextual and semantically aware recommendation
  7. Top-N recommendation
  8. Emerging themes
  9. Industry session 1
  10. Industry session 2
  11. Short papers
  12. Industry short papers
  13. Demonstrations
  14. Doctoral symposium
  15. Workshop outlines

Keynote address

Online controlled experiments: introduction, learnings, and humbling statistics BIBAFull-Text 1-2
  Ron Kohavi
The web provides an unprecedented opportunity to accelerate innovation by evaluating ideas quickly and accurately using controlled experiments (e.g., A/B tests and their generalizations). Whether for front-end user-interface changes, or backend recommendation systems and relevance algorithms, online controlled experiments are now utilized to make data-driven decisions at Amazon, Microsoft, eBay, Facebook, Google, Yahoo, Zynga, and at many other companies. While the theory of a controlled experiment is simple, and dates back to Sir Ronald A. Fisher's experiments at the Rothamsted Agricultural Experimental Station in England in the 1920s, the deployment and mining of online controlled experiments at scale -- thousands of experiments now -- has taught us many lessons. We provide an introduction, share real examples, key learnings, cultural challenges, and humbling statistics.

Tutorials

Conducting user experiments in recommender systems BIBAFull-Text 3-4
  Bart P. Knijnenburg
There is an increasing consensus in the field of recommender systems that we should move beyond the offline evaluation of algorithms towards a more user-centric approach. This tutorial teaches the essential skills involved in conducting user experiments, the scientific approach to user-centric evaluation. Such experiments are essential in uncovering how and why the user experience of recommender systems comes about.
Personality-based recommender systems: an overview BIBAFull-Text 5-6
  Maria Augusta S. N. Nunes; Rong Hu
Personality is a critical factor which influences people's behavior and interests. There is a high potential that incorporating users' characteristics into recommender systems could enhance recommendation quality and user experience. The goal of this tutorial is to give an overview of personality-based recommender systems and discuss challenges and possible research directions in this topic.
Building industrial-scale real-world recommender systems BIBAFull-Text 7-8
  Xavier Amatriain
In 2006, Netflix announced the Netflix Prize, a machine learning and data mining competition for movie rating prediction. We offered $1 million to whoever improved the accuracy of our existing system called Cinematch by 10%. We conducted this competition to find new ways to improve the recommendations we provide to our members, which is a key part of our business. However, we had to come up with a proxy question that was easier to evaluate and quantify: the root mean squared error (RMSE) of the predicted rating. A year into the competition, the Korbell team won the first Progress Prize with an 8.43% improvement. They reported more than 2000 hours of work in order to come up with the final combination of 107 algorithms that gave them this prize. And, they gave us the source code. We looked at the two underlying algorithms with the best performance in the ensemble. To put these algorithms to use, we had to work to overcome some limitations, for instance that they were built to handle 100 million ratings, instead of the more than 5 billion that we have, and that they were not built to adapt as members added more ratings. But once we overcame those challenges, we put the two algorithms into production, where they are still used as part of our recommendation engine.
   You might be wondering what happened with the final Grand Prize ensemble that won the $1M two years later. This is a truly impressive compilation and culmination of years of work, blending hundreds of predictive models to finally cross the finish line. We evaluated some of the new methods offline but the additional accuracy gains that we measured did not seem to justify the engineering effort needed to bring them into a production environment.
   This example highlights the fact that, besides improving offline metrics such as the RMSE, recommender systems need to take into account other practical issues such as scalability or deployment. In this tutorial, we go over some of those practical issues that many times are as important as the theory, if not more, in order to build an industrial-scale real-world recommender system.
The challenge of recommender systems challenges BIBAFull-Text 9-10
  Alan Said; Domonkos Tikk; Andreas Hotho
Recommender System Challenges such as the Netflix Prize, KDD Cup, etc. have contributed vastly to the development and adoptability of recommender systems. Each year a number of challenges or contests are organized covering different aspects of recommendation. In this tutorial and panel, we present some of the factors involved in successfully organizing a challenge, whether for reasons purely related to research, industrial challenges, or to widen the scope of recommender systems applications.

Multi-objective recommendation and human factors

Multiple objective optimization in recommender systems BIBAFull-Text 11-18
  Mario Rodriguez; Christian Posse; Ethan Zhang
We address the problem of optimizing recommender systems for multiple relevance objectives that are not necessarily aligned. Specifically, given a recommender system that optimizes for one aspect of relevance, semantic matching (as defined by any notion of similarity between source and target of recommendation; usually trained on CTR), we want to enhance the system with additional relevance signals that will increase the utility of the recommender system, but that may simultaneously sacrifice the quality of the semantic match. The issue is that semantic matching is only one relevance aspect of the utility function that drives the recommender system, albeit a significant aspect. In talent recommendation systems, job posters want candidates who are a good match to the job posted, but also prefer those candidates to be open to new opportunities. Recommender systems that recommend discussion groups must ensure that the groups are relevant to the users' interests, but also need to favor active groups over inactive ones. We refer to these additional relevance signals (job-seeking intent and group activity) as extraneous features, and they account for aspects of the utility function that are not captured by the semantic match (i.e. post-CTR down-stream utilities that reflect engagement: time spent reading, sharing, commenting, etc). We want to include these extraneous features into the recommendations, but we want to do so while satisfying the following requirements: 1) we do not want to drastically sacrifice the quality of the semantic match, and 2) we want to quantify exactly how the semantic match would be affected as we control the different aspects of the utility function. In this paper, we present an approach that satisfies these requirements.
   We frame our approach as a general constrained optimization problem and suggest ways in which it can be solved efficiently by drawing from recent research on optimizing non-smooth rank metrics for information retrieval. Our approach features the following characteristics: 1) it is model and feature agnostic, 2) it does not require additional labeled training data to be collected, and 3) it can be easily incorporated into an existing model as an additional stage in the computation pipeline. We validate our approach in a revenue-generating recommender system that ranks billions of candidate recommendations on a daily basis and show that a significant improvement in the utility of the recommender system can be achieved with an acceptable and predictable degradation in the semantic match quality of the recommendations.
Pareto-efficient hybridization for multi-objective recommender systems BIBAFull-Text 19-26
  Marco Tulio Ribeiro; Anisio Lacerda; Adriano Veloso; Nivio Ziviani
Performing accurate suggestions is an objective of paramount importance for effective recommender systems. Other important and increasingly evident objectives are novelty and diversity, which are achieved by recommender systems that are able to suggest diversified items not easily discovered by the users. Different recommendation algorithms have particular strengths and weaknesses when it comes to each of these objectives, motivating the construction of hybrid approaches. However, most of these approaches only focus on optimizing accuracy, with no regard for novelty and diversity. The problem of combining recommendation algorithms grows significantly harder when multiple objectives are considered simultaneously. For instance, devising multi-objective recommender systems that suggest items that are simultaneously accurate, novel and diversified may lead to a conflicting-objective problem, where the attempt to improve an objective further may result in worsening other competing objectives. In this paper we propose a hybrid recommendation approach that combines existing algorithms which differ in their level of accuracy, novelty and diversity. We employ an evolutionary search for hybrids following the Strength Pareto approach, which isolates hybrids that are not dominated by others (i.e., the so called Pareto frontier). Experimental results on two recommendation scenarios show that: (i) we can combine recommendation algorithms in order to improve an objective without significantly hurting other objectives, and (ii) we allow for adjusting the compromise between accuracy, diversity and novelty, so that the recommendation emphasis can be adjusted dynamically according to the needs of different users.
User effort vs. accuracy in rating-based elicitation BIBAFull-Text 27-34
  Paolo Cremonesi; Franca Garzottto; Roberto Turrin
One of the unresolved issues when designing a recommender system is the number of ratings -- i.e., the profile length -- that should be collected from a new user before providing recommendations. A design tension exists, induced by two conflicting requirements. On the one hand, the system must collect "enough" ratings from the user in order to learn her/his preferences and improve the accuracy of recommendations. On the other hand, gathering more ratings adds a burden on the user, which may negatively affect the user experience. Our research investigates the effects of profile length from both a subjective (user-centric) point of view and an objective (accuracy-based) perspective. We carried on an offline simulation with three algorithms, and a set of online experiments involving overall 960 users and four recommender algorithms, to measure which of the two contrasting forces influenced by the number of collected ratings -- recommendations relevance and burden of the rating process -- has stronger effects on the perceived quality of the user experience. Moreover, our study identifies the potentially optimal profile length for an explicit, rating based, and human controlled elicitation strategy.
TasteWeights: a visual interactive hybrid recommender system BIBAFull-Text 35-42
  Svetlin Bostandjiev; John O'Donovan; Tobias Höllerer
This paper presents an interactive hybrid recommendation system that generates item predictions from multiple social and semantic web resources, such as Wikipedia, Facebook, and Twitter. The system employs hybrid techniques from traditional recommender system literature, in addition to a novel interactive interface which serves to explain the recommendation process and elicit preferences from the end user. We present an evaluation that compares different interactive and non-interactive hybrid strategies for computing recommendations across diverse social and semantic web APIs. Results of the study indicate that explanation and interaction with a visual representation of the hybrid system increase user satisfaction and relevance of predicted content.

Social recommendation

Inspectability and control in social recommenders BIBAFull-Text 43-50
  Bart P. Knijnenburg; Svetlin Bostandjiev; John O'Donovan; Alfred Kobsa
Users of social recommender systems may want to inspect and control how their social relationships influence the recommendations they receive, especially since recommendations of social recommenders are based on friends rather than anonymous "nearest neighbors". We performed an online user experiment (N=267) with a Facebook music recommender system that gives users control over the recommendations, and explains how they came about. The results show that inspectability and control indeed increase users' perceived understanding of and control over the system, their rating of the recommendation quality, and their satisfaction with the system.
Spotting trends: the wisdom of the few BIBAFull-Text 51-58
  Xiaolan Sha; Daniele Quercia; Pietro Michiardi; Matteo Dell'Amico
Social media sites have used recommender systems to suggest items users might like but are not already familiar with. These items are typically movies, books, pictures, or songs. Here we consider an alternative class of items -- pictures posted by design-conscious individuals. We do so in the context of a mobile application in which users find "cool" items in the real world, take pictures of them, and share those pictures online. In this context, temporal dynamics matter, and users would greatly profit from ways of identifying the latest design trends. We propose a new way of recommending trending pictures to users, which unfolds in three steps. First, two types of users are identified -- those who are good at uploading trends (trend makers) and those who are experienced in discovering trends (trend spotters). Second, based on what those "special few" have uploaded and rated, trends are identified early on. Third, trends are recommended using existing algorithms. Upon the complete longitudinal dataset of the mobile application, we compare our approach's performance to a traditional recommender system's.
Real-time top-n recommendation in social streams BIBAFull-Text 59-66
  Ernesto Diaz-Aviles; Lucas Drumond; Lars Schmidt-Thieme; Wolfgang Nejdl
The Social Web is successfully established, and steadily growing in terms of users, content and services. People generate and consume data in real-time within social networking services, such as Twitter, and increasingly rely upon continuous streams of messages for real-time access to fresh knowledge about current affairs. In this paper, we focus on analyzing social streams in real-time for personalized topic recommendation and discovery. We consider collaborative filtering as an online ranking problem and present Stream Ranking Matrix Factorization -- RMFX -, which uses a pairwise approach to matrix factorization in order to optimize the personalized ranking of topics. Our novel approach follows a selective sampling strategy to perform online model updates based on active learning principles, that closely simulates the task of identifying relevant items from a pool of mostly uninteresting ones. RMFX is particularly suitable for large scale applications and experiments on the "476 million Twitter tweets" dataset show that our online approach largely outperforms recommendations based on Twitter's global trend, and it is also able to deliver highly competitive Top-N recommendations faster while using less space than Weighted Regularized Matrix Factorization (WRMF), a state-of-the-art matrix factorization technique for Collaborative Filtering, demonstrating the efficacy of our approach.
On top-k recommendation using social networks BIBAFull-Text 67-74
  Xiwang Yang; Harald Steck; Yang Guo; Yong Liu
Recommendation accuracy can be improved by incorporating trust relationships derived from social networks. Most recent work on social network based recommendation is focused on minimizing the root mean square error (RMSE). Social network based top-k recommendation, which recommends to a user a small number of items at a time, is not well studied. In this paper, we conduct a comprehensive study on improving the accuracy of top-k recommendation using social networks. We first show that the existing social-trust enhanced Matrix Factorization (MF) models can be tailored for top-k recommendation by including observed and missing ratings in their training objective functions. We also propose a Nearest Neighbor (NN) based top-k recommendation method that combines users' neighborhoods in the trust network with their neighborhoods in the latent feature space. Experimental results on two publicly available datasets show that social networks can significantly improve the top-k hit ratio, especially for cold start users. Surprisingly, we also found that the technical approach for combining feedback data (e.g. ratings) with social network information that works best for minimizing RMSE works poorly for maximizing the hit ratio, and vice versa.

Implicit feedback and user preference

Optimal radio channel recommendations with explicit and implicit feedback BIBAFull-Text 75-82
  Omar Moling; Linas Baltrunas; Francesco Ricci
The very large majority of recommender systems are running as server-side applications, and they are controlled by the content provider, i.e., who provides the recommended items. This paper focuses on a different scenario: the user is supposed to be able to access content from multiple providers, in our application they offer radio channels, and it is up to a personal recommender installed on the clients' side to decide which channel to select and recommend to the user. We exploit the implicit feedback derived from the user's listening behavior, and we model channel recommendation as a sequential decision making problem. We have implemented a personal RS that integrates reinforcement learning techniques to decide what channel to play every time the user asks for a new music track or the current track finishes playing. In a live user study we show that the proposed system can sequentially select the next channel to play such that the users listen to the streamed tracks for a larger fraction, and for more time, compared to a baseline system not exploiting implicit feedback.
Alternating least squares for personalized ranking BIBAFull-Text 83-90
  Gábor Takács; Domonkos Tikk
Two flavors of the recommendation problem are the explicit and the implicit feedback settings. In the explicit feedback case, users rate items and the user-item preference relationship can be modelled on the basis of the ratings. In the harder but more common implicit feedback case, the system has to infer user preferences from indirect information: presence or absence of events, such as a user viewed an item. One approach for handling implicit feedback is to minimize a ranking objective function instead of the conventional prediction mean squared error. The naive minimization of a ranking objective function is typically expensive. This difficulty is usually overcome by a trade-off: sacrificing the accuracy to some extent for computational efficiency by sampling the objective function. In this paper, we present a computationally effective approach for the direct minimization of a ranking objective function, without sampling. We demonstrate by experiments on the Y!Music and Netflix data sets that the proposed method outperforms other implicit feedback recommenders in many cases in terms of the ErrorRate, ARP and Recall evaluation metrics.
Local implicit feedback mining for music recommendation BIBAFull-Text 91-98
  Diyi Yang; Tianqi Chen; Weinan Zhang; Qiuxia Lu; Yong Yu
Digital music has experienced a quite fascinating transformation during the past decades. Thousands of people share or distribute their music collections on the Internet, resulting in an explosive increase of information and more user dependence on automatic recommender systems. Though there are many techniques such as collaborative filtering, most approaches focus mainly on users' global behaviors, neglecting local actions and the specific properties of music. In this paper, we propose a simple and effective local implicit feedback model mining users' local preferences to get better recommendation performance in both rating and ranking prediction. Moreover, we design an efficient training algorithm to speed up the updating procedure, and give a method to find the most appropriate time granularity to assist the performance. We conduct various experiments to evaluate the performance of this model, which show that it outperforms baseline model significantly. Integration with existing temporal models achieves a great improvement compared to the reported best single model for Yahoo! Music.
How many bits per rating? BIBAFull-Text 99-106
  Daniel Kluver; Tien T. Nguyen; Michael Ekstrand; Shilad Sen; John Riedl
Most recommender systems assume user ratings accurately represent user preferences. However, prior research shows that user ratings are imperfect and noisy. Moreover, this noise limits the measurable predictive power of any recommender system. We propose an information theoretic framework for quantifying the preference information contained in ratings and predictions. We computationally explore the properties of our model and apply our framework to estimate the efficiency of different rating scales for real world datasets. We then estimate how the amount of information predictions give to users is related to the scale ratings are collected on. Our findings suggest a tradeoff in rating scale granularity: while previous research indicates that coarse scales (such as thumbs up / thumbs down) take less time, we find that ratings with these scales provide less predictive value to users. We introduce a new measure, preference bits per second, to quantitatively reconcile this tradeoff.

Contextual and semantically aware recommendation

High quality recommendations for small communities: the case of a regional parent network BIBAFull-Text 107-114
  Sven Strickroth; Niels Pinkwart
Traditional recommender systems are well established in scenarios in which "enough" items, users and ratings are available for the algorithms to operate on. However, automatic recommendations are also desirable in smaller online communities which only contain several hundred items and users. Collaborative filters, as one of the most successful technologies for recommender systems, do not perform well here. This paper argues that recommender systems can make use of contextual information and domain specific semantics in order to be able to generate recommendations also for these smaller usage scenarios. The new hybrid recommendation approach presented in the paper enhances traditional neighborhood-based collaborative filtering techniques through the use of new kinds of data and a combination of different recommendation methods (rule, demographic, and average based). While the algorithmic techniques presented in this paper are suitable (especially) for smaller online communities, they can also be applied to improve the quality of recommendations in larger communities. The approach was implemented and evaluated in a small regional bound parent education community. A multi-staged evaluation was conducted in order to determine the quality of recommendations: A cross-validation (recall), an expert questionnaire (recommendation quality) and a field study (user satisfaction). The results show that recommenders even for smaller communities are possible and can produce high quality recommendations.
Finding a needle in a haystack of reviews: cold start context-based hotel recommender system BIBAFull-Text 115-122
  Asher Levi; Osnat Mokryn; Christophe Diot; Nina Taft
Online hotel searching is a daunting task due to the wealth of online information. Reviews written by other travelers replace the word-of-mouth, yet turn the search into a time consuming task. Users do not rate enough hotels to enable a collaborative filtering based recommendation. Thus, a cold start recommender system is needed. In this work we design a cold start hotel recommender system, which uses the text of the reviews as its main data. We define context groups based on reviews extracted from TripAdvisor.com and Venere.com. We introduce a novel weighted algorithm for text mining. Our algorithm imitates a user that favors reviews written with the same trip intent and from people of similar background (nationality) and with similar preferences for hotel aspects, which are our defined context groups. Our approach combines numerous elements, including unsupervised clustering to build a vocabulary for hotel aspects, semantic analysis to understand sentiment towards hotel features, and the profiling of intent and nationality groups.
   We implemented our system which was used by the public to conduct 150 trip planning experiments. We compare our solution to the top suggestions of the mentioned web services and show that users were, on average, 20% more satisfied with our hotel recommendations. We outperform these web services even more in cities where hotel prices are high.
Review quality aware collaborative filtering BIBAFull-Text 123-130
  Sindhu Raghavan; Suriya Gunasekar; Joydeep Ghosh
Probabilistic matrix factorization (PMF) and other popular approaches to collaborative filtering assume that the ratings given by users for products are genuine, and hence they give equal importance to all available ratings. However, this is not always true due to several reasons including the presence of opinion spam in product reviews. In this paper, the possibility of performing collaborative filtering while attaching weights or quality scores to the ratings is explored. The quality scores, which are determined from the corresponding review data are used to "up-weight" or "down-weight" the importance given to the individual rating while performing collaborative filtering, thereby improving the accuracy of the predictions. First, the measure used to capture the quality of the ratings is described. Different approaches for estimating the quality score based on the available review information are examined. Subsequently, a mathematical formulation to incorporate quality scores as weights for the ratings in the basic PMF framework is derived. Experimental evaluation on two product categories of a benchmark data set from Amazon.com demonstrates the efficacy of our approach.
Context-aware music recommendation based on latent topic sequential patterns BIBAFull-Text 131-138
  Negar Hariri; Bamshad Mobasher; Robin Burke
Contextual factors can greatly influence the users' preferences in listening to music. Although it is hard to capture these factors directly, it is possible to see their effects on the sequence of songs liked by the user in his/her current interaction with the system. In this paper, we present a context-aware music recommender system which infers contextual information based on the most recent sequence of songs liked by the user. Our approach mines the top frequent tags for songs from social tagging Web sites and uses topic modeling to determine a set of latent topics for each song, representing different contexts. Using a database of human-compiled playlists, each playlist is mapped into a sequence of topics and frequent sequential patterns are discovered among these topics. These patterns represent frequent sequences of transitions between the latent topics representing contexts. Given a sequence of songs in a user's current interaction, the discovered patterns are used to predict the next topic in the playlist. The predicted topics are then used to post-filter the initial ranking produced by a traditional recommendation algorithm. Our experimental evaluation suggests that our system can help produce better recommendations in comparison to a conventional recommender system based on collaborative or content-based filtering. Furthermore, the topic modeling approach proposed here is also useful in providing better insight into the underlying reasons for song selection and in applications such as playlist construction and context prediction.

Top-N recommendation

CLiMF: learning to maximize reciprocal rank with collaborative less-is-more filtering BIBAFull-Text 139-146
  Yue Shi; Alexandros Karatzoglou; Linas Baltrunas; Martha Larson; Nuria Oliver; Alan Hanjalic
In this paper we tackle the problem of recommendation in the scenarios with binary relevance data, when only a few (k) items are recommended to individual users. Past work on Collaborative Filtering (CF) has either not addressed the ranking problem for binary relevance datasets, or not specifically focused on improving top-k recommendations. To solve the problem we propose a new CF approach, Collaborative Less-is-More Filtering (CLiMF). In CLiMF the model parameters are learned by directly maximizing the Mean Reciprocal Rank (MRR), which is a well-known information retrieval metric for measuring the performance of top-k recommendations. We achieve linear computational complexity by introducing a lower bound of the smoothed reciprocal rank metric. Experiments on two social network datasets demonstrate the effectiveness and the scalability of CLiMF, and show that CLiMF significantly outperforms a naive baseline and two state-of-the-art CF methods.
Ranking with non-random missing ratings: influence of popularity and positivity on evaluation metrics BIBAFull-Text 147-154
  Bruno Pradel; Nicolas Usunier; Patrick Gallinari
The evaluation of recommender systems in terms of ranking has recently gained attention, as it seems to better fit the top-k recommendation task than the usual ratings prediction task. In that context, several authors have proposed to consider missing ratings as some form of negative feedback to compensate for the skewed distribution of observed ratings when users choose the items they rate. In this work, we study two major biases of the selection of items: the first one is that some items obtain more ratings than others (popularity effect), and the second one is that positive ratings are observed more frequently than negative ratings (positivity effect). We present a theoretical analysis and experiments on the Yahoo! dataset with randomly selected items, which show that considering missing data as a form of negative feedback during training may improve performances, but also that it can be misleading when testing, favoring models of popularity more than models of user preferences.
Sparse linear methods with side information for top-n recommendations BIBAFull-Text 155-162
  Xia Ning; George Karypis
The increasing amount of side information associated with the items in E-commerce applications has provided a very rich source of information that, once properly exploited and incorporated, can significantly improve the performance of the conventional recommender systems. This paper focuses on developing effective algorithms that utilize item side information for top-N recommender systems. A set of sparse linear methods with side information (SSLIM) is proposed, which involve a regularized optimization process to learn a sparse aggregation coefficient matrix based on both user-item purchase profiles and item side information. This aggregation coefficient matrix is used within an item-based recommendation framework to generate recommendations for the users. Our experimental results demonstrate that SSLIM outperforms other methods in effectively utilizing side information and achieving performance improvement.
Scalable similarity-based neighborhood methods with MapReduce BIBAFull-Text 163-170
  Sebastian Schelter; Christoph Boden; Volker Markl
Similarity-based neighborhood methods, a simple and popular approach to collaborative filtering, infer their predictions by finding users with similar taste or items that have been similarly rated. If the number of users grows to millions, the standard approach of sequentially examining each item and looking at all interacting users does not scale. To solve this problem, we develop a MapReduce algorithm for the pairwise item comparison and top-N recommendation problem that scales linearly with respect to a growing number of users. This parallel algorithm is able to work on partitioned data and is general in that it supports a wide range of similarity measures. We evaluate our algorithm on a large dataset consisting of 700 million song ratings from Yahoo! Music.

Emerging themes

An approach to context-based recommendation in software development BIBAFull-Text 171-178
  Bruno Antunes; Joel Cordeiro; Paulo Gomes
A software developer programming in an object-oriented programming language deals with a source code structure that may contain hundreds of source code elements. These elements are commonly related to each other and working on a specific element may require the developer to access other related elements. We propose a recommendation approach that uses the context of the developer to retrieve and rank recommendations of relevant source code elements in the IDE. These recommendations provide a shortcut to reach the desired elements and increase the awareness of the developer in relation to elements that may be of interest in that moment. We have tested our approach with a group of developers and the results show that context has a promising role in predicting and ranking the source code elements needed by a developer at each moment.
A semantic approach to recommending text advertisements for images BIBAFull-Text 179-186
  Weinan Zhang; Li Tian; Xinruo Sun; Haofen Wang; Yong Yu
In recent years, more and more images have been uploaded and published on the Web. Along with text Web pages, images have been becoming important media to place relevant advertisements. Visual contextual advertising, a young research area, refers to finding relevant text advertisements for a target image without any textual information (e.g., tags). There are two existing approaches, advertisement search based on image annotation, and more recently, advertisement matching based on feature translation between images and texts. However, the state of the art fails to achieve satisfactory results due to the fact that recommended advertisements are syntactically matched but semantically mismatched. In this paper, we propose a semantic approach to improving the performance of visual contextual advertising. More specifically, we exploit a large high-quality image knowledge base (ImageNet) and a widely-used text knowledge base (Wikipedia) to build a bridge between target images and advertisements. The image-advertisement match is built by mapping images and advertisements into the respective knowledge bases and then finding semantic matches between the two knowledge bases. The experimental results show that semantic match outperforms syntactic match significantly using test images from Flickr. We also show that our approach gives a large improvement of 16.4% on the precision of the top 10 matches over previous work, with more semantically relevant advertisements recommended.
Ads and the city: considering geographic distance goes a long way BIBAFull-Text 187-194
  Diego Saez-Trumper; Daniele Quercia; Jon Crowcroft
Social-networking sites have started to offer tools that suggest "guests" who should be invited to user-defined social events (e.g., birthday parties, networking events). The problem of how to recommend people to events is similar to the more traditional (recommender system) problem of how to recommend events (items) to people (users). Yet, upon Foursquare data of "who visits what" in the city of London, we show that a state-of-the-art recommender system does not perform well -- mainly because of data sparsity. To fix this problem, we add domain knowledge to the recommendation process. From the complex system literature in human mobility, we learn two insights: 1) there are special individuals (often called power users) who visit many places; and 2) individuals go to a venue not only because they like it but also because they are close-by. We model these insights into two simple models and learn that: 1) simply recommending power users works better than random but is far from producing the best recommendations; 2) an item-based recommender system produces accurate recommendations; and 3) recommending places that are closest to a user's geographic center of interest produces recommendations that are as accurate as, if not more accurate than, item-based recommender's. This last result has practical implications as it offers guidelines for designing location-based recommender systems and for partly addressing cold-start situations.
BlurMe: inferring and obfuscating user gender based on ratings BIBAFull-Text 195-202
  Udi Weinsberg; Smriti Bhagat; Stratis Ioannidis; Nina Taft
User demographics, such as age, gender and ethnicity, are routinely used for targeting content and advertising products to users. Similarly, recommender systems utilize user demographics for personalizing recommendations and overcoming the cold-start problem. Often, privacy-concerned users do not provide these details in their online profiles. In this work, we show that a recommender system can infer the gender of a user with high accuracy, based solely on the ratings provided by users (without additional metadata), and a relatively small number of users who share their demographics. Focusing on gender, we design techniques for effectively adding ratings to a user's profile for obfuscating the user's gender, while having an insignificant effect on the recommendations provided to that user.

Industry session 1

Distributed, real-time Bayesian learning in online services BIBAFull-Text 203-204
  Ralf Herbrich
The last ten years have seen a tremendous growth in Internet-based online services such as search, advertising, gaming and social networking. Today, it is important to analyze large collections of user interaction data as a first step in building predictive models for these services as well as learn these models in real-time. One of the biggest challenges in this setting is scale: not only does the sheer scale of data necessitate parallel processing but it also necessitates distributed models; with over 900 million active users at Facebook, any user-specific sets of features in a linear or non-linear model yields models of a size bigger than can be stored in a single system.
   In this talk, I will give a hands-on introduction to one of the most versatile tools for handling large collections of data with distributed probabilistic models: the sum-product algorithm for approximate message passing in factor graphs. I will discuss the application of this algorithm for the specific case of generalized linear models and outline the challenges of both approximate and distributed message passing including an in-depth discussion of expectation propagation and Map-Reduce.
Recommendation challenges in web media settings BIBAFull-Text 205-206
  Ronny Lempel
This paper calls out several research challenges in the art of recommendation technology as applied in Web media sites. One particular characteristic of such recommendation settings is the relative low cost of falsely recommending an irrelevant item, which means that recommendation schemes can be less conservative and more exploratory. This also creates opportunities for better item cold-start handling. Other technical difficulties include analyzing offline data that is heavily biased by the site's appearance, and in a related vein -- once a recommendation module's appearance has been designed -- defining the correct metrics by which to measure it. Also called out are tradeoffs between personalization and contextualization, as are novel schemes that aim at recommending sets and sequences of items.

Industry session 2

I've got 10 million songs in my pocket: now what? BIBAFull-Text 207-208
  Paul B. Lamere
The proverbial celestial jukebox has become a reality. With today's online music services a music fan is never more than a few clicks away from being able to listen to nearly any song that has ever been recorded. Recommender systems can play a key role in this new music ecosystem, helping listeners explore, discover, organize and share music. However, in many ways music recommendation is very different than recommendation in other well-studied domains such as books and movies. In this talk we explore how recommender systems can be used in the music space, and the particular challenges that the music domain presents to the designers of recommender systems.

Short papers

Dynamic personalized recommendation of comment-eliciting stories BIBAFull-Text 209-212
  Michal Aharon; Amit Kagian; Ronny Lempel; Yehuda Koren
Media Websites often solicit users' comments on content items such as videos, news stories, blog posts, etc. Commenting activity increases user engagement with the sites, by both comment writers and readers, and so sites are looking for ways to increase the volume of comments. This work develops a recommender system aiming to present users with items -- news stories, in our case -- on which they are likely to comment. We combine items' content with a collaborative-filtering approach (utilizing users' co-commenting patterns) in a latent factor modeling framework. Building upon previous work, we focus on a continuous, real-time approach to address the problem above. After an initial training period during which commenting activity of users is observed, the system is tested at each subsequent comment submission event by predicting which story is being commented on by a given user at a given time. Our results show that we are able to overcome the site's inherent presentation bias and outperform a strong baseline as users' commenting history grows.
Using graph partitioning techniques for neighbour selection in user-based collaborative filtering BIBAFull-Text 213-216
  Alejandro Bellogin; Javier Parapar
Spectral clustering techniques have become one of the most popular clustering algorithms, mainly because of their simplicity and effectiveness. In this work, we make use of one of these techniques, Normalised Cut, in order to derive a cluster-based collaborative filtering algorithm which outperforms other standard techniques in the state-of-the-art in terms of ranking precision. We frame this technique as a method for neighbour selection, and we show its effectiveness when compared with other cluster-based methods. Furthermore, the performance of our method could be improved if standard similarity metrics -- such as Pearson's correlation -- are also used when predicting the user's preferences.
Remembering the stars?: effect of time on preference retrieval from memory BIBAFull-Text 217-220
  Dirk Bollen; Mark Graus; Martijn C. Willemsen
Many recommendation systems rely on explicit ratings provided by their users. Often these ratings are provided long after consuming the item, relying heavily on people's representation of the quality of the item in memory. This paper investigates a psychological process, the "positivity effect", that influences the retrieval of quality judgments from our memory by which pleasant items are being processed and recalled from memory more effectively than unpleasant items. In an offline study on the MovieLens data we used the time between release date and rating date as a proxy for the time between consumption and rating. Ratings for movies tend to increase over time, consistent with the positivity effect. A subsequent online user study used a direct measure of time between rating and consumption, by asking users to rate movies (recently aired on television) and to explicitly report how long ago they watched these movies. In contrast to the offline study we find that ratings tend to decline over time showing reduced accuracy in ratings for items experienced long ago. We discuss the impact these rating dynamics might have on recommender algorithms, especially in cases where a new user has to submit his preferences to a system.
Local learning of item dissimilarity using content and link structure BIBAFull-Text 221-224
  Abir De; Maunendra Sankar Desarkar; Niloy Ganguly; Pabitra Mitra
In the Recommendation Problem, it is often important to find a set of items similar to a particular item or a group of items. This problem of finding similar items for the recommendation task may also be viewed as a link prediction problem in a network, where the items can be treated as the nodes. The strength of the edge connecting two items represents the similarity between the items. In this context, a central challenge is to suitably define an appropriate dissimilarity function between the items. For content based recommender systems, the dissimilarity function should take into account the individual attributes of the items. The same attribute may have different importances in different parts of the underlying network. We focus on the problem of learning a suitable dissimilarity function between items and address it by formulating it as a constrained optimization problem which captures the local weightages of the attributes in different regions of the graph. The constraints are imposed in such a way that the non-connected nodes show higher value of dissimilarity than the connected nodes. The local tuning of the weights learns the optimal value of weights in various parts of the network: from the portions having rich graph information to the portions having only content information. Detailed experimentation shows the superiority of the proposed algorithm over the Adamic Adar metric as well as logistic regression methodology.
Design and evaluation of a group recommender system BIBAFull-Text 225-228
  Toon De Pessemier; Simon Dooms; Luc Martens
Though most recommender systems make suggestions for individual users, in many circumstances the selected items (e.g., movies) are not for personal usage but rather for consumption in group. In this paper, we present a recommender system for audio-visual content that generates suggestions for groups of people (such as families or friends) in the home environment. In this context, different group recommendation strategies are evaluated for various algorithms and sizes of the group. An offline evaluation proves the assumption that for randomly composed groups the accuracy of all recommendation algorithms decreases if the group size grows. Besides, the results show that the group recommendation strategy which produces the most accurate results is depending on the algorithm that is used for generating individual recommendations. Consequently, if an existing recommender system for individuals is extended to a recommender system for groups, the group recommendation strategy has to be chosen based on the utilized recommendation algorithm in order to maximize the efficiency of the group recommendations.
Swarming to rank for recommender systems BIBAFull-Text 229-232
  Ernesto Diaz-Aviles; Mihai Georgescu; Wolfgang Nejdl
Recommender systems make product suggestions that are tailored to the user's individual needs and represent powerful means to combat information overload. In this paper, we focus on the item prediction task of Recommender Systems and present SwarmRankCF, a method to automatically optimize the performance quality of recommender systems using a Swarm Intelligence perspective. Our approach, which is well-founded in a Particle Swarm Optimization framework, learns a ranking function by optimizing the combination of unique characteristics (i.e., features) of users, items and their interactions. In particular, we build feature vectors from a factorization of the user-item interaction matrix, and directly optimize Mean Average Precision metric in order to learn a linear ranking model for personalized recommendations. Our experimental evaluation, on a real world online radio dataset, indicates that our approach is able to find ranking functions that significantly improve the performance of the system for the Top-N recommendation task.
When recommenders fail: predicting recommender failure for algorithm selection and combination BIBAFull-Text 233-236
  Michael Ekstrand; John Riedl
Hybrid recommender systems -- systems using multiple algorithms together to improve recommendation quality -- have been well-known for many years and have shown good performance in recent demonstrations such as the NetFlix Prize. Modern hybridization techniques, such as feature-weighted linear stacking, take advantage of the hypothesis that the relative performance of recommenders varies by circumstance and attempt to optimize each item score to maximize the strengths of the component recommenders. Less attention, however, has been paid to understanding what these strengths and failure modes are. Understanding what causes particular recommenders to fail will facilitate better selection of the component recommenders for future hybrid systems and a better understanding of how individual recommender personalities can be harnessed to improve the recommender user experience. We present an analysis of the predictions made by several well-known recommender algorithms on the MovieLens 10M data set, showing that for many cases in which one algorithm fails, there is another that will correctly predict the rating.
Constrained collective matrix factorization BIBAFull-Text 237-240
  Yu-Jia Huang; Evan Wei Xiang; Rong Pan
Transfer learning for collaborative filtering (TLCF) aims to solve the sparsity problem by transferring rating knowledge across multiple domains. Taking domain difference into account, one of the issues in cross-domain collaborative filtering is to selectively transfer knowledge from source/auxiliary domains. In particular, this paper addresses the problem of inconstant users (users with changeable preferences across different domains) when transferring knowledge about users from another auxiliary domain. We first formulate the problem of inconstant users caused by domain difference and then propose a new model that performs constrained collective matrix factorization (CCMF). Our experiments on simulated and real data show that CCMF has superior performance than other methods.
Recommending academic papers via users' reading purposes BIBAFull-Text 241-244
  Yichen Jiang; Aixia Jia; Yansong Feng; Dongyan Zhao
The past decades have witnessed the rapid development of academic research, which results in a growing number of scholarly papers. As a result, paper recommender systems have been proposed to help researchers find their interested papers. Most previous studies in paper recommendations mainly concentrate on paper-paper or user-paper similarities without taking users' reading purposes into account. It is common that different users may prefer to different aspects of a paper, e.g., the focused problem/task or the proposed solution. In this paper, we propose to satisfy user-specific reading purposes by recommending the most problem-related papers or solution-related papers to users separately. For a target paper, we use the paper citation graph to generate a set of potential relevant papers. Once getting the candidate set, we calculate the problem-based similarities and solution-based similarities between candidates and the target paper through a concept based topic model, respectively. We evaluate our models on a real academic paper dataset and our experiments show that our approach outperforms a traditional similarity based model and can provide highly relevant paper recommendations according to different reading purposes for researchers.
Influential seed items recommendation BIBAFull-Text 245-248
  Qi Liu; Biao Xiang; Enhong Chen; Yong Ge; Hui Xiong; Tengfei Bao; Yi Zheng
In this paper, we present a systematic perspective study on choosing and evaluating the initial seed items that will be recommended to the cold start users. We first construct an item consumption correlation network to capture the existing users' general consumption behaviors. Then, we formalize initial items recommendation as the influential seed set selection problem. Along this line, we present several methods, each of which selects seed items according to different rules. Finally, the experimental results on two real-world data sets verify that with different seed items, the users' consumption numbers will be quite different. Meanwhile, the results also provide many deep insights into these selection methods and their recommended seed items.
Discovering latent factors from movies genres for enhanced recommendation BIBAFull-Text 249-252
  Marcelo G. Manzato
Current approaches on collaborative filtering factorize user-item matrices in order to infer latent factors from ratings previously assigned by users. However, they all have to deal with sparseness, whose workarounds are prone to bias and/or overfitting. This paper proposes a recommender algorithm that is based on a factorized matrix composed of user preferences associated to the movies' genres/categories. The advantage of using such user-genre matrix factorization model is that it requires less computational resources, as the matrix will be less sparse and at lower dimension. We present the experimental results with a dataset composed of real users, comparing the performance of different modules of our algorithm.
Exploiting the web of data in model-based recommender systems BIBAFull-Text 253-256
  Tommaso Di Noia; Roberto Mirizzi; Vito Claudio Ostuni; Davide Romito
The availability of a huge amount of interconnected data in the so called Web of Data (WoD) paves the way to a new generation of applications able to exploit the information encoded in it. In this paper we present a model-based recommender system leveraging the datasets publicly available in the Linked Open Data (LOD) cloud as DBpedia and LinkedMDB. The proposed approach adapts support vector machine (SVM) to deal with RDF triples. We tested our system and showed its effectiveness by a comparison with different recommender systems techniques -- both content-based and collaborative filtering ones.
Probabilistic news recommender systems with feedback BIBAFull-Text 257-260
  Shankar Prawesh; Balaji Padmanabhan
In prior work we addressed a major problem faced by media sites with popularity based recommender systems such as the top-10 list of most liked or most clicked posts. We showed that the hard cutoff used in these systems to generate the "Top N" lists is prone to unduly penalizing good articles that may have just missed the cutoff. A solution to this was to generate recommendations probabilistically, which is an approach that has been shown to be robust against some manipulation techniques as well. The aim of this research is to introduce a class of probabilistic news recommender systems that incorporates widely practiced recommendation techniques as a special case. We establish our results in a special case of two articles using the urn models with feedback mechanism from probability theory.
Collaborative learning of preference rankings BIBAFull-Text 261-264
  Tim Salimans; Ulrich Paquet; Thore Graepel
We propose a model for learning user preference rankings for the purpose of making product recommendations. The model allows us to learn from pairwise preference statements or from (incomplete) rankings over more than two items. We present two algorithms for performing inference in this model, both with excellent scaling in the number of users and items. The superior predictive performance of the new method is demonstrated on the well-known sushi preference data set. In addition, we show how the model can be used effectively in an active learning setting where we select only a small number of informative items for learning.
Making recommendations in a microblog to improve the impact of a focal user BIBAFull-Text 265-268
  Shanchan Wu; Leanna Gong; William Rand; Louiqa Raschid
We present a microblog recommendation system that can help monitor users, track conversations, and potentially improve diffusion impact. Given a Twitter network of active users and their followers, and historical activity of tweets, retweets and mentions, we build upon a prediction tool to predict the Top K users who will retweet or mention a focal user, in the future [10]. We develop personalized recommendations for each focal user. We identify characteristics of focal users such as the size of the follower network, or the level of sentiment averaged over all tweets; both have an impact on the quality of personalized recommendations. We use (high) betweenness centrality as a proxy of attractive users to target when making recommendations. Our recommendations successfully identify a greater fraction of users with higher betweenness centrality, in comparison to the overall distribution of betweenness centrality of the ground truth users for some focal user.
The influence of knowledgeable explanations on users' perception of a recommender system BIBAFull-Text 269-272
  Markus Zanker
Recommender Systems (RS) help online customers in identifying those items from a variety of choices that best match their presumed needs and preferences. In this context explanations summarize the reasons why a specific item is proposed and are capable of increasing the users' trust in the system's results. This paper presents results from an online experiment on a real-world platform indicating that explanations are an essential piece of functionality of a recommendation system, that significantly increases users' perception of the utility of a recommender system, the intention to use it repeatedly as well as the commitment to recommend it to others.

Industry short papers

Social referral: leveraging network connections to deliver recommendations BIBAFull-Text 273-276
  Mohammad Shafkat Amin; Baoshi Yan; Sripad Sriram; Anmol Bhasin; Christian Posse
Much work has been done to study the interplay between recommender systems and social networks. This creates a very powerful coupling in presenting highly relevant recommendations to the users. However, to our knowledge, little attention has been paid to leverage a user's social network to deliver these recommendations. We present a novel approach to aid delivery of recommendations using the recipient's friends or connections. Our contributions with this study are 1) A novel recommendation delivery paradigm called Social Referral, which utilizes a user's social network for the delivery of relevant content. 2) An implementation of the paradigm is described in a real industrial production setting of a large online professional network. 3) A study of the interaction between the trifecta of the recommender system, the trusted connections and the end consumer of the recommendation by comparing and contrasting the proposed approach's performance with the direct recommender system.
   Our experiments indicate that Social Referral is a promising mechanism for recommendation delivery. The experiments show that a significant portion of users are receptive to passing along relevant recommendations to their social networks, and that recommendations delivered through users' social networks are much more likely to be accepted than those directly delivered to users.
Case study on the business value impact of personalized recommendations on a large online retailer BIBAFull-Text 277-280
  Thiago Belluf; Leopoldo Xavier; Ricardo Giglio
While much of the recommender systems literature has focused on the off-line evaluation of prediction performance, a few case studies using online controlled experiments that assess the performance of business indicators are available. In this article, we describe the methods and results of an ongoing investigation conducted on the business value impact of personalized recommendations on three different portals of Nova Pontocom, the second largest Latin American online retailer. An on-line controlled experiment (A/B testing), conducted for one month and covering 600,000 distinct users, statistically points out to a general revenue increase in the order of 8-20%. In addition, other consumer behavior metrics such as the number of page views and the more diverse distribution of sales among the products catalog also support the positive impact of personalized recommendations in terms of business value.
The Xbox recommender system BIBAFull-Text 281-284
  Noam Koenigstein; Nir Nice; Ulrich Paquet; Nir Schleyen
A recent addition to Microsoft's Xbox Live Marketplace is a recommender system which allows users to explore both movies and games in a personalized context. The system largely relies on implicit feedback, and runs on a large scale, serving tens of millions of daily users. We describe the system design, and review the core recommendation algorithm.
Enlister: baidu's recommender system for the biggest Chinese Q&A website BIBAFull-Text 285-288
  Qiwen Liu; Tianjian Chen; Jing Cai; Dianhai Yu
In this paper, we describe the concept & design of a real-time question RS (recommender system), the Enlister project, for the biggest Chinese Q&A (Questions and Answers) website and evaluate its performance on massive data from this real-world practice. We demonstrate how we weigh in among different recommendation algorithms and optimization methods. To enhance recommendation accuracy and handling time-sensitive questions, we propose a large scale real-time RS based on the combination of machine learning algorithms and the stream computing technology. Considering of algorithm flexibility and performance, we use the maximum entropy model as the fundamental model design in the CTR (click-through rate) prediction of recommendation items. In the perspective of the Enlister system architecture, we illustrate how we divide and conquer massive data processing problem with a novel stream computing design which reduces the data process latency down to seconds. Finally we analyze the online test result and prove our design concept by achieving a series of significant improvements.
HeyStaks: a real-world deployment of social search BIBAFull-Text 289-292
  Barry Smyth; Maurice Coyle; Peter Briggs
The purpose of this paper is to provide a deployment update for the HeyStaks social search system which uses recommendation techniques to add collaboration to mainstream search engines such as Google, Bing, and Yahoo. We describe our the results of initial deployments, including an assessment of the quality of HeyStaks' recommendations, and highlight some lessons learned in the marketplace.

Demonstrations

A system for Twitter user list curation BIBAFull-Text 293-294
  Igor Brigadir; Derek Greene; Padraig Cunningham
With increased adoption of social networking tools, it is becoming more difficult to extract useful information from the mass of data generated daily by users. Curation of content and sources is an important filter in separating the signal from noise. A good set of credible sources often requires painstaking manual curation, which often yields incomplete coverage of a topic. In this demo, we present a recommender system to aid this process, improving the quality and quantity of sources. The system is highly-adaptable to the goals of the curator, enabling some novel uses for curating and monitoring lists of users.
CubeThat: news article recommender BIBAFull-Text 295-296
  Sidharth Chhabra; Paul Resnick
The CubeThat browser extension for Chrome displays recommended additional news stories related to the same topic as the current news story. The recommended stories are organized into clusters, and clusters that the user has already sampled from are grayed out, in order to encourage users to explore multiple aspects of a story. Users can also provide feedback to improve the clustering, by dragging stories from one cluster to another.
The demonstration of the reviewer's assistant BIBAFull-Text 297-298
  Ruihai Dong; Markus Schaal; Michael P. O'Mahony; Kevin McCarthy; Barry Smyth
User generated reviews are now a familiar and valuable part of most e-commerce sites since high quality reviews are known to influence purchasing decisions. In this demonstration we describe work on the Reviewer's Assistant (RA), which is a recommendation system that is designed to help users to write better quality reviews. It does this by suggesting relevant topics that they may wish to discuss based on the product they are reviewing and the content of their review so far.
Recommenders for the enterprise: event, contact, and group BIBAFull-Text 299-300
  Abigail Gertner; Beth Lavender; James Winston
This extended abstract provides a summary of MITRE's Recommender demonstrations for ACM RecSys 2012. We have three demonstrations: Event Recommender, Contact Recommender, and Group Recommender. MITRE is a technology company with thousands of staff and varied government customers. Help from recommenders is becoming a necessity to find the right resources to bring to bear on large, multidisciplinary problems. We have developed and deployed these recommenders to explore their effectiveness, implementation challenges for an enterprise, and their use models.
Integrated content marketing BIBAFull-Text 301-302
  James Griffin
idio is a technology startup that works with major brands in both the UK and the US to drive Customer Centricity through Content Marketing. idio's full-stack platform fuses Natural Language Processing, Data Mining, Web Analytics and Recommendation Engines at scale.
Using ratings to profile your health BIBAFull-Text 303-304
  Neal Lathia
The widespread adoption of mobile technology allows personalised applications to be deployed in an increasing host of contexts; user modelling, profiling, and personalised recommendations are becoming an integral component of mobile information systems. Furthermore, mobile technology enables the recording and collection of facets of daily life, which has given rise to the notion of the quantified self; researchers operating at the intersection of computer and social science are now seeking to understand how these mobiles' data can aide the design of health interventions and inform future psychological and social science research. In this work, we describe the design of a personalised mobile application that seeks user feedback and builds a user profile about people's gastrointestinal health using ratings and tags. We describe the application's design and the personalised health insights it provides (and, particularly, why recommendations were not designed as a means for self-diagnosis).
Finding a needle in a haystack of reviews: cold start context-based hotel recommender system demo BIBAFull-Text 305-306
  Asher Levi; Osnat Mokryn; Christophe Diot; Nina Taft
Online hotel searching is a daunting task due to the wealth of online information. Reviews written by other travelers replace the word-of-mouth, yet turn the search into a time consuming task. Users do not rate enough hotels to enable a collaborative filtering based recommendation. Thus, a cold start recommender system is needed. This demo describes briefly our cold start hotel recommender system, which uses the text of the reviews as its main data. We define context groups based on reviews extracted from TripAdvisor.com and Venere.com. We introduce a novel weighted algorithm for text mining.
   We implemented our system which was used by the public to conduct 150 trip planning experiments. We compare our solution to the top suggestions of the mentioned web services and show that users were, on average, 20% more satisfied with our hotel recommendations. We outperform these web services even more in cities where hotel prices are high.
Yokie: explorations in curated real-time search & discovery using Twitter BIBAFull-Text 307-308
  Owen Phelan; Kevin McCarthy; Barry Smyth
Our research involves developing technology and techniques that apply the vast sea of real-time web data to interesting problems and topics. In this demo, we will present the ongoing development of a novel real-time search and discovery service named Yokie (http://yok.ie, early technology description originally published in [1]). Yokie uses the large volume of hyperlink-laden messages on social networks like Twitter as the basis of its content and ranking systems. Curated sets of users (or "Search Parties") form the basis of sourcing the content from the networks, and the metadata of the containing messages form the basis of ranking and contextual retrieval of the hyperlinks. Each hyperlink is indexed with a compound set of terms from multiple tweets (should the given hyperlink be shared more than once). This indexing step is a novel example of collaborative tagging of resources. The application is live with more than 100 users, who have performed approximately 1000 queries. We will demonstrate the main techniques and novel ranking and retrieval techniques and user features.
pGPA: a personalized grade prediction tool to aid student success BIBAFull-Text 309-310
  Mark Sheehan; Young Park
Many educational institutions are starting to make use of their scholastic data to improve the academic experience for their students. To aid in this endeavor we have developed a research prototype implementation of a collaborative filtering-based tool called the personalized Grade Prediction Advisor (pGPA). The goal of this prototype tool is to demonstrate the potential of recommender technology by providing grade predictions for upcoming courses in a student's academic career to support decision-making for administrators, students, educators, and academic advisors. In this demonstration we briefly describe the underlying technology and potential applications of pGPA. We then present how a user can interact with pGPA to produce and interpret personalized grade predictions for an individual student or group of students.
Recommending interesting events in real-time with foursquare check-ins BIBAFull-Text 311-312
  Max Sklar; Blake Shaw; Andrew Hogue
Foursquare is a location-based social application that helps users explore the world around them and share their experiences with friends. When foursquare users visit places, they "check in" using their mobile phones, indicating they are at that place. People check in for a variety of reasons: to keep up with friends, get tips about places, redeem rewards, and keep track of their personal history. In aggregate, billions of these check-ins reveal distinct patterns about when places are popular and allow us to build a unique place recommendation engine which can identify and recommend interesting events in real-time based on statistical deviations from past historical trends.

Doctoral symposium

An open framework for multi-source, cross-domain personalisation with semantic interest graphs BIBAFull-Text 313-316
  Benjamin Heitmann
Cross-domain recommendations are currently available in closed, proprietary social networking ecosystems such as Facebook, Twitter and Google+. I propose an open framework as an alternative, which enables cross-domain recommendations with domain-agnostic user profiles modeled as semantic interest graphs. This novel framework covers all parts of a recommender system. It includes an architecture for privacy-enabled profile exchange, a distributed and domain-agnostic user model and a cross-domain recommendation algorithm. This enables users to receive recommendations for a target domain (e.g. food) based on any kind of previous interests.
Exploiting the characteristics of matrix factorization for active learning in recommender systems BIBAFull-Text 317-320
  Rasoul Karimi; Christoph Freudenthaler; Alexandros Nanopoulos; Lars Schmidt-Thieme
Recommender systems help web users to address information overload. However their performance depends on the number of provided ratings by users. This problem is amplified for a new user because he/she has not provided any rating. To address this problem, active learning methods have been proposed to acquire those ratings from users, that will help most in determining their interests. However, different from the classic active learning, users (the "oracle") are not always able to provide an answer for queries. The easiest way to solve this problem is to ask most popular items, i.e items which have received many ratings from training users. But it is static and presents the same items to all users regardless of the ratings they have provided so far. In this paper we propose a method that improves the most popular selection strategy using the characteristics of matrix factorization. It finds similar users to the new user in the latent space and then selects item which is most popular among the similar users. The experimental results show the proposed method outperforms the most popular method both in terms of error and the number of received ratings.
Dynamically selecting an appropriate context type for personalisation BIBAFull-Text 321-324
  Tomaš Kramár; Mária Bieliková
Narrowing down the context in the ranking phase of information retrieval has been shown to produce results that are more relevant to searcher's need. We have identified two types of contexts that could be used in the process of personalisation. We research these contexts in the domain of personalised search, but show that our approach can be used for any kind of personalisation or recommendation. We focus on two aspects of the context: temporal context and activity-based context and describe a more general personalisation framework based on lightweight semantics, that can leverage any type of context.
Utilising document content for tag recommendation in folksonomies BIBAFull-Text 325-328
  Nikolas Landia
Real-world tagging datasets have a large proportion of new/unseen documents. Few approaches for recommending tags to a user for a document address this new item problem, concentrating instead on artificially created post-core datasets where it is guaranteed that the user as well as the document of each test post is known to the system and already has some tags assigned to it. In order to recommend tags for unseen documents, approaches are required which model documents not only based on the tags assigned to it in the past (if any), but also the content. The focus of my research is on utilising the content of documents in order to address the new item problem in tag recommendation. I apply this methodology first to simple baseline tag recommenders and then the more advanced tag recommendation algorithm FolkRank.
   One of my main contributions is a novel adaptation to the FolkRank graph model to use multiple word nodes instead of a single document node to represent each document. This enables FolkRank to recommend tags for unseen documents and makes it applicable to full real-world tagging datasets, addressing the new item problem in tag recommendation.
Using group recommendation heuristics for the prioritization of requirements BIBAFull-Text 329-332
  Gerald Ninaus
Group recommendation heuristics have been successfully applied in different domains such as Interactive Television and e-Tourism. Our work focuses on the improvement of requirements prioritization techniques on the basis of group recommendation technologies. First, we analyse the impact of preference visibility on the outcome of the requirements prioritization process. In this context we evaluate the resulting software quality and the satisfaction of stakeholders with the requirements prioritization. Second, we analyse standard group recommendation heuristics on a dataset originated from a study conducted at our university. Finally, we propose new heuristics to improve the prediction quality and conclude with an evaluation of the different recommendation heuristics.
Beyond lists: studying the effect of different recommendation visualizations BIBAFull-Text 333-336
  Denis Parra
Recommendation Systems have been studied from several perspectives over the last twenty years -- prediction accuracy, algorithmic scalability, knowledge sources, types of recommended items and tasks, evaluation methods, etc. -- but one area that has not been deeply investigated is the effect of different visualizations and their interaction with personal traits on users' evaluation of the recommended items. In this paper, I survey visual approaches that go beyond presenting the recommended items as a textual list or as annotations in context. I also review related literature from recommendations' explanations. In this thesis, I aim to understand how different visualizations and some personal traits might influence users' assessment of recommended items, particularly in domains where multidimensional data or contextual constraints are involved. I present the prototype of 2 recommendation visualizations and then briefly propose the research approach of this investigation.
The user-centered design of a recommender system for a universal library catalogue BIBAFull-Text 337-340
  Simon Wakeling
This paper describes a collaborative project between the University of Sheffield's iSchool and OCLC (an international library cooperative), the aim of which is to develop a prototype recommender system for WorldCat.org, the aggregated catalogue of OCLC's member libraries. This paper describes a user-centered approach, utilizing both qualitative and quantitative methods, which aims to establish how and why users engage with library catalogues and WorldCat.org in particular, whether there is a need for recommendations in the library domain, and if so what type of recommendations best support the information-seeking needs of users. An outline of the proposed methodology is provided, along with a report on work completed to date. An analysis of UK library catalogues shows the prevalence of recommender systems to be very low, while initial results from focus group interviews and a pop-up survey show a significant demand for recommendations from two key user-groups (students and academics).
Reducing the sparsity of contextual information for recommender systems BIBAFull-Text 341-344
  Dusan Zelenik; Maria Bielikova
Our work focuses on the improvement of the accuracy of context-aware recommender systems. Contextual information showed to be promising factor in recommender systems. However, pure context-based recommender systems can not outperform other approaches mainly due to high sparsity of contextual information. We propose an idea to improve accuracy of context based recommender systems by context inference. Context inference is based on effect discovered by analyses of the context as a factor influencing user needs. Analyses of the news readers reveals existence of behavioural correlation which is the main pillar of proposed context inference. Method for context inference is based on collaborative filtering and clustering of web usage (as a non-discretizing alternative to association rules mining).

Workshop outlines

4th ACM RecSys workshop on recommender systems and the social web BIBFull-Text 345-346
  Bamshad Mobasher; Dietmar Jannach; Werner Geyer; Andreas Hotho
RecSys'12 workshop on human decision making in recommender systems BIBAFull-Text 347-348
  Marco de Gemmis; Alexander Felfernig; Pasquale Lops; Francesco Ricci; Giovanni Semeraro; Martijn C. Willemsen
Interacting with a recommender system means to take different decisions such as selecting an item from a recommendation list, selecting a specific item feature value (e.g., camera's size, zoom) as a search criteria, selecting feedback features to be critiqued in a critiquing based recommendation session, or selecting a repair proposal for inconsistent user preferences when interacting with a knowledge-based recommender. In all these situations, users face a decision task. This workshop (Decisions@RecSys) focuses on approaches for supporting effective and efficient human decision making in different types of recommendation scenarios.
4th workshop on context-aware recommender systems (CARS 2012) BIBAFull-Text 349-350
  Gediminas Adomavicius; Linas Baltrunas; Ernesto William de Luca; Tim Hussein; Alexander Tuzhilin
CARS 2012 builds upon the success of the three previous editions held in conjunction with the 3rd to 5th ACM Conferences on Recommender Systems from 2009 to 2011. The 1st CARS Workshop was held in New York, NY, USA, whereas Barcelona, Spain, was home of the 2nd CARS Workshop in 2010. In 2011, the 3rd CARS workshop was held in Chicago, IL, USA.
Workshop on recommendation utility evaluation: beyond RMSE -- RUE 2012 BIBAFull-Text 351-352
  Xavier Amatriain; Pablo Castells; Arjen de Vries; Christian Posse
Measuring the error in rating prediction has been by far the dominant evaluation methodology in the Recommender Systems literature. Yet there seems to be a general consensus that this criterion alone is far from being enough to assess the practical effectiveness of a recommender system. Information Retrieval metrics have started to be used to evaluate item selection and ranking rather than rating prediction, but considerable divergence remains in the adoption of such metrics by different authors. On the other hand, recommendation utility includes other key dimensions and concerns beyond accuracy, such as novelty and diversity, user engagement, and business performance. While the need for further extension, formalization, clarification and standardization of evaluation methodologies is recognized in the community, this need is still unmet for a large extent. The RUE 2012 workshop sought to identify and better understand the current gaps in recommender system evaluation methodologies, help lay directions for progress in addressing them, and contribute to the consolidation and convergence of experimental methods and practice.
Recommender systems challenge 2012 BIBAFull-Text 353-354
  Nikos Manouselis; Alan Said; Domonkos Tikk; Jannis Hermanns; Benjamin Kille; Hendrik Drachsler; Katrien Verbert; Kris Jack
The Recommender System Challenge 2012 invited participants to work on two tracks with real-world datasets and to submit their contributions that would be related to specific problem contexts. First of all, it asked participants to develop new algorithms and to compare them to other algorithms in given settings; in addition, it asked participants to explore with new recommendation methods, services, as well as added-value services related to recommendation.
RecSys'12 workshop on interfaces for recommender systems (InterfaceRS'12) BIBFull-Text 355-356
  Nava Tintarev; Rong Hu; Pearl Pu
1st workshop on recommendation technologies for lifestyle change 2012 BIBAFull-Text 357-358
  Bernd Ludwig; Francesco Ricci; Zerrin Yumak
The workshop on Recommendation Technologies for Lifestyle Change will be an opportunity for discussing open issues, and propose technical solutions for the designing of intelligent information systems that can support and promote lifestyle change. The objective of these systems is to provide users with up-to-date information, and help them to make choices in every day life activities establishing a sustainable compromise between quality of life, individuality, and fun.
Personalizing the local mobile experience: workshop at RecSys 2012 BIBAFull-Text 359-360
  Henriette Cramer; Karen Church; Neal Lathia; Daniele Quercia
Mobile, local recommendations are on the rise. Surprisingly however, research addressing user perceptions of local recommendations and local differences when interacting with such recommendation services is yet scarce. Location-based recommendation services are mostly evaluated from a "ecommendation systems" standpoint, with limited experiential insights from users and limited focus on local differences that may apply. This workshop focuses on the local, personal user experience, and provides a forum to exchange experiences, insights and strategies in personalizing local mobile applications and generating local recommendations that fit local user needs.