{"title":"Opening Remarks","authors":"Shilad Sen, Werner Geyer, J. Freyne, P. Castells","doi":"10.1145/2959100.3057279","DOIUrl":"https://doi.org/10.1145/2959100.3057279","url":null,"abstract":"","PeriodicalId":315651,"journal":{"name":"Proceedings of the 10th ACM Conference on Recommender Systems","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122610205","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
We introduce an asynchronous distributed stochastic gradient algorithm for matrix factorization based collaborative filtering. The main idea of this approach is to distribute the user-rating matrix across different machines, each having access only to a part of the information, and to asynchronously propagate the updates of the stochastic gradient optimization across the network. Each time a machine receives a parameter vector, it averages its current parameter vector with the received one, and continues its iterations from this new point. Additionally, we introduce a similarity based regularization that constrains the user and item factors to be close to the average factors of their similar users and items found on subparts of the distributed user-rating matrix. We analyze the impact of the regularization terms on MovieLens (100K, 1M, 10M) and NetFlix datasets and show that it leads to a more efficient matrix factorization in terms of Root Mean Square Error (RMSE) and Mean Absolute Error (MAE), and that the asynchronous distributed approach significantly improves in convergence time as compared to an equivalent synchronous distributed approach.
{"title":"Asynchronous Distributed Matrix Factorization with Similar User and Item Based Regularization","authors":"Bikash Joshi, F. Iutzeler, Massih-Reza Amini","doi":"10.1145/2959100.2959161","DOIUrl":"https://doi.org/10.1145/2959100.2959161","url":null,"abstract":"We introduce an asynchronous distributed stochastic gradient algorithm for matrix factorization based collaborative filtering. The main idea of this approach is to distribute the user-rating matrix across different machines, each having access only to a part of the information, and to asynchronously propagate the updates of the stochastic gradient optimization across the network. Each time a machine receives a parameter vector, it averages its current parameter vector with the received one, and continues its iterations from this new point. Additionally, we introduce a similarity based regularization that constrains the user and item factors to be close to the average factors of their similar users and items found on subparts of the distributed user-rating matrix. We analyze the impact of the regularization terms on MovieLens (100K, 1M, 10M) and NetFlix datasets and show that it leads to a more efficient matrix factorization in terms of Root Mean Square Error (RMSE) and Mean Absolute Error (MAE), and that the asynchronous distributed approach significantly improves in convergence time as compared to an equivalent synchronous distributed approach.","PeriodicalId":315651,"journal":{"name":"Proceedings of the 10th ACM Conference on Recommender Systems","volume":"181 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126172327","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In 2006, Netflix announced a $1M prize competition to advance recommendation algorithms. The recommendation problem was simplified as the accuracy in predicting a user rating measured by the Root Mean Squared Error. While that formulation helped get the attention of the research community, it put the focus on the wrong approach and metric while leaving many important factors out. In this tutorial we will describe the advances in Recommender Systems in the last 10 years from an industry perspective based on the instructors' personal experience at companies like Quora, LinkedIn, Netflix, or Yahoo! We will do so in the form of different lessons learned through the years. Some of those lessons will describe the different components of modern recommender systems such as: personalized ranking, similarity, explanations, context-awareness, or multi-armed bandits. Others will also review the usage of novel algorithmic approaches such as Factorization Machines, Restricted Boltzmann Machines, SimRank, Deep Neural Networks, or Listwise Learning-to-rank. Others will dive into details of the importance of gathering the right data or using the correct optimization metric. But, most importantly, we will give many examples of prototypical industrial-scale recommender systems with special focus on those unsolved challenges that should define the future of the recommender systems area.
{"title":"Tutorial: Lessons Learned from Building Real-life Recommender Systems","authors":"X. Amatriain, D. Agarwal","doi":"10.1145/2959100.2959194","DOIUrl":"https://doi.org/10.1145/2959100.2959194","url":null,"abstract":"In 2006, Netflix announced a $1M prize competition to advance recommendation algorithms. The recommendation problem was simplified as the accuracy in predicting a user rating measured by the Root Mean Squared Error. While that formulation helped get the attention of the research community, it put the focus on the wrong approach and metric while leaving many important factors out. In this tutorial we will describe the advances in Recommender Systems in the last 10 years from an industry perspective based on the instructors' personal experience at companies like Quora, LinkedIn, Netflix, or Yahoo! We will do so in the form of different lessons learned through the years. Some of those lessons will describe the different components of modern recommender systems such as: personalized ranking, similarity, explanations, context-awareness, or multi-armed bandits. Others will also review the usage of novel algorithmic approaches such as Factorization Machines, Restricted Boltzmann Machines, SimRank, Deep Neural Networks, or Listwise Learning-to-rank. Others will dive into details of the importance of gathering the right data or using the correct optimization metric. But, most importantly, we will give many examples of prototypical industrial-scale recommender systems with special focus on those unsolved challenges that should define the future of the recommender systems area.","PeriodicalId":315651,"journal":{"name":"Proceedings of the 10th ACM Conference on Recommender Systems","volume":"140 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116512613","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Determinantal point processes (DPPs) are an emerging model for encoding probabilities over subsets, such as shopping baskets, selected from a ground set, such as an item catalog. They have recently proved to be appealing models for a number of machine learning tasks, including product recommendation. DPPs are parametrized by a positive semi-definite kernel matrix. Prior work has shown that using a low-rank factorization of this kernel provides scalability improvements that open the door to training on large-scale datasets and computing online recommendations, both of which are infeasible with standard DPP models that use a full-rank kernel. A low-rank DPP model can be trained using an optimization-based method, such as stochastic gradient ascent, to find a point estimate of the kernel parameters, which can be performed efficiently on large-scale datasets. However, this approach requires careful tuning of regularization parameters to prevent overfitting and provide good predictive performance, which can be computationally expensive. In this paper we present a Bayesian method for learning a low-rank factorization of this kernel, which provides automatic control of regularization. We show that our Bayesian low-rank DPP model can be trained efficiently using stochastic gradient Hamiltonian Monte Carlo (SGHMC). Our Bayesian model generally provides better predictive performance on several real-world product recommendation datasets than optimization-based low-rank DPP models trained using stochastic gradient ascent, and better performance than several state-of-the art recommendation methods in many cases.
{"title":"Bayesian Low-Rank Determinantal Point Processes","authors":"Mike Gartrell, U. Paquet, Noam Koenigstein","doi":"10.1145/2959100.2959178","DOIUrl":"https://doi.org/10.1145/2959100.2959178","url":null,"abstract":"Determinantal point processes (DPPs) are an emerging model for encoding probabilities over subsets, such as shopping baskets, selected from a ground set, such as an item catalog. They have recently proved to be appealing models for a number of machine learning tasks, including product recommendation. DPPs are parametrized by a positive semi-definite kernel matrix. Prior work has shown that using a low-rank factorization of this kernel provides scalability improvements that open the door to training on large-scale datasets and computing online recommendations, both of which are infeasible with standard DPP models that use a full-rank kernel. A low-rank DPP model can be trained using an optimization-based method, such as stochastic gradient ascent, to find a point estimate of the kernel parameters, which can be performed efficiently on large-scale datasets. However, this approach requires careful tuning of regularization parameters to prevent overfitting and provide good predictive performance, which can be computationally expensive. In this paper we present a Bayesian method for learning a low-rank factorization of this kernel, which provides automatic control of regularization. We show that our Bayesian low-rank DPP model can be trained efficiently using stochastic gradient Hamiltonian Monte Carlo (SGHMC). Our Bayesian model generally provides better predictive performance on several real-world product recommendation datasets than optimization-based low-rank DPP models trained using stochastic gradient ascent, and better performance than several state-of-the art recommendation methods in many cases.","PeriodicalId":315651,"journal":{"name":"Proceedings of the 10th ACM Conference on Recommender Systems","volume":"47 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115908101","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Item-based approaches based on SLIM (Sparse LInear Methods) have demonstrated very good performance for top-N recommendation; however they only estimate a single model for all the users. This work is based on the intuition that not all users behave in the same way -- instead there exist subsets of like-minded users. By using different item-item models for these user subsets, we can capture differences in their preferences and this can lead to improved performance for top-N recommendations. In this work, we extend SLIM by combining global and local SLIM models. We present a method that computes the prediction scores as a user-specific combination of the predictions derived by a global and local item-item models. We present an approach in which the global model, the local models, their user-specific combination, and the assignment of users to the local models are jointly optimized to improve the top-N recommendation performance. Our experiments show that the proposed method improves upon the standard SLIM model and outperforms competing top-N recommendation approaches.
{"title":"Local Item-Item Models For Top-N Recommendation","authors":"Evangelia Christakopoulou, G. Karypis","doi":"10.1145/2959100.2959185","DOIUrl":"https://doi.org/10.1145/2959100.2959185","url":null,"abstract":"Item-based approaches based on SLIM (Sparse LInear Methods) have demonstrated very good performance for top-N recommendation; however they only estimate a single model for all the users. This work is based on the intuition that not all users behave in the same way -- instead there exist subsets of like-minded users. By using different item-item models for these user subsets, we can capture differences in their preferences and this can lead to improved performance for top-N recommendations. In this work, we extend SLIM by combining global and local SLIM models. We present a method that computes the prediction scores as a user-specific combination of the predictions derived by a global and local item-item models. We present an approach in which the global model, the local models, their user-specific combination, and the assignment of users to the local models are jointly optimized to improve the top-N recommendation performance. Our experiments show that the proposed method improves upon the standard SLIM model and outperforms competing top-N recommendation approaches.","PeriodicalId":315651,"journal":{"name":"Proceedings of the 10th ACM Conference on Recommender Systems","volume":"70 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132349005","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
At Quora, our mission is to share and grow the world's knowledge. Recommender systems are at the core of this mission: we need to recommend the most important questions to people most likely to write great answers, and recommend the best answers to people interested in reading them. Driven by the above mission statement, we have a variety of interesting and challenging recommendation problems and a large, rich data set that we can work with to build novel solutions for them. In this talk, we will describe several of these recommendation problems and present our approaches solving them.
{"title":"Recommending the World's Knowledge: Application of Recommender Systems at Quora","authors":"Lei Yang, X. Amatriain","doi":"10.1145/2959100.2959128","DOIUrl":"https://doi.org/10.1145/2959100.2959128","url":null,"abstract":"At Quora, our mission is to share and grow the world's knowledge. Recommender systems are at the core of this mission: we need to recommend the most important questions to people most likely to write great answers, and recommend the best answers to people interested in reading them. Driven by the above mission statement, we have a variety of interesting and challenging recommendation problems and a large, rich data set that we can work with to build novel solutions for them. In this talk, we will describe several of these recommendation problems and present our approaches solving them.","PeriodicalId":315651,"journal":{"name":"Proceedings of the 10th ACM Conference on Recommender Systems","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133723080","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In the past decade, news consumption through traditional channels such as print has been on the decline while online and digital news consumption has been steadily growing. Bloomberg, renowned for its products in the financial world, has a very strong presence in the news and media industry. Bloomberg Media, on an average, publishes 400-500 stories and videos per day and we have close to 30 million unique visitors on our websites and mobile applications every month consuming this content. At such a scale it is very important to recommend relevant information for a good user experience. Recommendations in the News and Media domain bring a unique set of challenges due to the dynamic nature of the data as well as unique consumption patterns. The biggest challenge with building recommendation systems in the News domain is the dynamic nature of the domain itself; new content is published every few minutes and majority of the content has a short shelf life, i.e., the news is not relevant to users after a certain time span and the time span is generally of the order of hours rather than days, making it important to deliver relevant content in a timely manner. Moreover, our users consume content differently based on time of day. For example, some users whose focus is market news and market data during the day consume more long form and generic articles and videos in the evening. User preferences, along with being cyclical in nature, tend to change over time, so algorithms need to adapt to the changing taste of the user. In addition, we need to ensure that the users do get their share of important/trending news and are not put into a filter bubble. In this talk, we will present some novel techniques we have applied to popular approaches in the field of Recommender Systems to be able to address the unique challenges which the news domain presents.
{"title":"News Recommendations at scale at Bloomberg Media: Challenges and Approaches","authors":"Dhaval Shah, Pramod Koneru, Parth Shah, Rohit Parimi","doi":"10.1145/2959100.2959118","DOIUrl":"https://doi.org/10.1145/2959100.2959118","url":null,"abstract":"In the past decade, news consumption through traditional channels such as print has been on the decline while online and digital news consumption has been steadily growing. Bloomberg, renowned for its products in the financial world, has a very strong presence in the news and media industry. Bloomberg Media, on an average, publishes 400-500 stories and videos per day and we have close to 30 million unique visitors on our websites and mobile applications every month consuming this content. At such a scale it is very important to recommend relevant information for a good user experience. Recommendations in the News and Media domain bring a unique set of challenges due to the dynamic nature of the data as well as unique consumption patterns. The biggest challenge with building recommendation systems in the News domain is the dynamic nature of the domain itself; new content is published every few minutes and majority of the content has a short shelf life, i.e., the news is not relevant to users after a certain time span and the time span is generally of the order of hours rather than days, making it important to deliver relevant content in a timely manner. Moreover, our users consume content differently based on time of day. For example, some users whose focus is market news and market data during the day consume more long form and generic articles and videos in the evening. User preferences, along with being cyclical in nature, tend to change over time, so algorithms need to adapt to the changing taste of the user. In addition, we need to ensure that the users do get their share of important/trending news and are not put into a filter bubble. In this talk, we will present some novel techniques we have applied to popular approaches in the field of Recommender Systems to be able to address the unique challenges which the news domain presents.","PeriodicalId":315651,"journal":{"name":"Proceedings of the 10th ACM Conference on Recommender Systems","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115694154","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Alexandros Karatzoglou, Balázs Hidasi, D. Tikk, Oren Sar Shalom, Haggai Roitman, Bracha Shapira, L. Rokach
We believe that Deep Learning is one of the next big things in Recommendation Systems technology. The past few years have seen the tremendous success of deep neural networks in a number of complex tasks such as computer vision, natural language processing and speech recognition. Despite this, only little work has been published on Deep Learning methods for Recommender Systems. Notable recent application areas are music recommendation, news recommendation, and session-based recommendation. The aim of the workshop is to encourage the application of Deep Learning techniques in Recommender Systems, to promote research in deep learning methods for Recommender Systems, and to bring together researchers from the Recommender Systems and Deep Learning communities.
{"title":"RecSys'16 Workshop on Deep Learning for Recommender Systems (DLRS)","authors":"Alexandros Karatzoglou, Balázs Hidasi, D. Tikk, Oren Sar Shalom, Haggai Roitman, Bracha Shapira, L. Rokach","doi":"10.1145/2959100.2959202","DOIUrl":"https://doi.org/10.1145/2959100.2959202","url":null,"abstract":"We believe that Deep Learning is one of the next big things in Recommendation Systems technology. The past few years have seen the tremendous success of deep neural networks in a number of complex tasks such as computer vision, natural language processing and speech recognition. Despite this, only little work has been published on Deep Learning methods for Recommender Systems. Notable recent application areas are music recommendation, news recommendation, and session-based recommendation. The aim of the workshop is to encourage the application of Deep Learning techniques in Recommender Systems, to promote research in deep learning methods for Recommender Systems, and to bring together researchers from the Recommender Systems and Deep Learning communities.","PeriodicalId":315651,"journal":{"name":"Proceedings of the 10th ACM Conference on Recommender Systems","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115712137","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Machine learning and recommendations systems have changed the way we interact with not just the internet, but some of the basic services that we use to organize and run our life. As the people that build these systems, we have a social responsibility to consider how these systems affect people, and furthermore, we should do whatever we can to prevent these models from perpetuating some of the prejudice and bias that exist in our society today. This talk will cover some of the recommendation systems that have gone wrong across various industries, and attempt to provide some solutions for raising awareness and prevention. Approaches that will be explored include using interpretable models, using ensemble models to separate features that shouldn't interact, and designing test data sets for capturing accidental bias.
{"title":"When Recommendation Systems Go Bad","authors":"Evan Estola","doi":"10.1145/2959100.2959117","DOIUrl":"https://doi.org/10.1145/2959100.2959117","url":null,"abstract":"Machine learning and recommendations systems have changed the way we interact with not just the internet, but some of the basic services that we use to organize and run our life. As the people that build these systems, we have a social responsibility to consider how these systems affect people, and furthermore, we should do whatever we can to prevent these models from perpetuating some of the prejudice and bias that exist in our society today. This talk will cover some of the recommendation systems that have gone wrong across various industries, and attempt to provide some solutions for raising awareness and prevention. Approaches that will be explored include using interpretable models, using ensemble models to separate features that shouldn't interact, and designing test data sets for capturing accidental bias.","PeriodicalId":315651,"journal":{"name":"Proceedings of the 10th ACM Conference on Recommender Systems","volume":"62 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116024919","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
How can you discover something new, that matches your interest? Recommender Systems have been studied since the 90ies. Their benefit comes from guiding a user through the density of the information jungle to useful knowledge clearings. Early research on recommender systems focuses on algorithms and their evaluation to improve recommendation accuracy using F-measures and other methodologies from signal-detection theory. Present research includes other aspects such as human factors that affect the user experience and interactive visualization techniques to support transparency of results and user control. In this paper, we analyze all publications on recommender systems from the scopus database, and particularly also papers with such an HCI focus. Based on an analysis of these papers, future topics for recommender systems research are identified, which include more advanced support for user control, adaptive interfaces, affective computing and applications in high risk domains.
{"title":"HCI for Recommender Systems: the Past, the Present and the Future","authors":"André Calero Valdez, M. Ziefle, K. Verbert","doi":"10.1145/2959100.2959158","DOIUrl":"https://doi.org/10.1145/2959100.2959158","url":null,"abstract":"How can you discover something new, that matches your interest? Recommender Systems have been studied since the 90ies. Their benefit comes from guiding a user through the density of the information jungle to useful knowledge clearings. Early research on recommender systems focuses on algorithms and their evaluation to improve recommendation accuracy using F-measures and other methodologies from signal-detection theory. Present research includes other aspects such as human factors that affect the user experience and interactive visualization techniques to support transparency of results and user control. In this paper, we analyze all publications on recommender systems from the scopus database, and particularly also papers with such an HCI focus. Based on an analysis of these papers, future topics for recommender systems research are identified, which include more advanced support for user control, adaptive interfaces, affective computing and applications in high risk domains.","PeriodicalId":315651,"journal":{"name":"Proceedings of the 10th ACM Conference on Recommender Systems","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129368656","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：