AutoRec: Autoencoders Meet Collaborative Filtering

Suvash Sedhain; Aditya Krishna Menon; Scott Sanner; Lexing Xie

doi:10.1145/2740908.2742726

AutoRec: Autoencoders Meet Collaborative Filtering

Published on May 18, 2015 in WWW (The Web Conference)

· DOI :10.1145/2740908.2742726

Suvash Sedhain5

Estimated H-index: 5

(ANU: Australian National University), Aditya Krishna Menon22

Estimated H-index: 22

(ANU: Australian National University)+ 1 AuthorsLexing Xie32

Estimated H-index: 32

(ANU: Australian National University)

Sources

Abstract

This paper proposes AutoRec, a novel autoencoder framework for collaborative filtering (CF). Empirically, AutoRec's compact and efficiently trainable model outperforms state-of-the-art CF techniques (biased matrix factorization, RBM-CF and LLORMA) on the Movielens and Netflix datasets.

Figures & Tables

📖 Papers frequently viewed together

Factorization meets the neighborhood: a multifaceted collaborative filtering model

2008KDD: Knowledge Discovery and Data Mining

1 Author (Yehuda Koren)

2,236 Citations

Deep Collaborative Filtering via Marginalized Denoising Auto-encoder

2015CIKM: Conference on Information and Knowledge Management

3 Authors (Sheng Li, ..., Yun Fu)

195 Citations

Item-based collaborative filtering recommendation algorithms

2001WWW: The Web Conference

4 Authors (Badrul M. Sarwar, ..., John Riedl)

5,487 Citations

References5

Newest

Local Low-Rank Matrix Approximation

Jun 16, 2013 in ICML (International Conference on Machine Learning)

#1Joonseok Lee (Georgia Institute of Technology)H-Index: 13

#2Seungyeon Kim (Georgia Institute of Technology)H-Index: 8

Last. Yoram Singer (Google)H-Index: 13

view all 4 authors...

Matrix approximation is a common tool in recommendation systems, text mining, and computer vision. A prevalent assumption in constructing matrix approximations is that the partially observed matrix is of low-rank. We propose a new matrix approximation model where we assume instead that the matrix is locally of low-rank, leading to a representation of the observed matrix as a weighted sum of low-rank matrices. We analyze the accuracy of the proposed local low-rank modeling. Our experiments show i...more

107 Citations

Matrix Factorization Techniques for Recommender Systems

Aug 1, 2009·IEEE Computer4.42

#1Yehuda Koren (Yahoo!)H-Index: 46

#2Robert M. BellH-Index: 53

Last. Chris VolinskyH-Index: 25

view all 3 authors...

As the Netflix Prize competition has demonstrated, matrix factorization models are superior to classic nearest neighbor techniques for producing product recommendations, allowing the incorporation of additional information such as implicit feedback, temporal effects, and confidence levels.

7,470 Citations Source

Restricted Boltzmann machines for collaborative filtering

Jun 20, 2007 in ICML (International Conference on Machine Learning)

#1Ruslan Salakhutdinov (U of T: University of Toronto)H-Index: 84

#2Andriy Mnih (U of T: University of Toronto)H-Index: 27

Last. Geoffrey E. Hinton (U of T: University of Toronto)H-Index: 132

view all 3 authors...

Most of the existing approaches to collaborative filtering cannot handle very large data sets. In this paper we show how a class of two-layer undirected graphical models, called Restricted Boltzmann Machines (RBM's), can be used to model tabular data, such as user's ratings of movies. We present efficient learning and inference procedures for this class of models and demonstrate that RBM's can be successfully applied to the Netflix data set, containing over 100 million user/movie ratings. We als...more

1,172 Citations Source

Item-based collaborative filtering recommendation algorithms

May 1, 2001 in WWW (The Web Conference)

#1Badrul M. Sarwar (UMN: University of Minnesota)H-Index: 9

#2George Karypis (UMN: University of Minnesota)H-Index: 77

Last. John Riedl (UMN: University of Minnesota)H-Index: 59

view all 4 authors...

Recommender systems apply knowledge discovery techniques to the problem of making personalized recommendations for information, products or services during a live interaction. These systems, especially the k-nearest neighbor collaborative ltering based ones, are achieving widespread success on the Web. The tremendous growth in the amount of available information and the number of visitors to Web sites in recent years poses some key challenges for recommender systems. These are: producing high qu...more

5,487 Citations Source

A direct adaptive method for faster backpropagation learning: the RPROP algorithm

#1Martin Riedmiller (KIT: Karlsruhe Institute of Technology)H-Index: 40

#2Heinrich Braun (KIT: Karlsruhe Institute of Technology)H-Index: 6

A learning algorithm for multilayer feedforward networks, RPROP (resilient propagation), is proposed. To overcome the inherent disadvantages of pure gradient-descent, RPROP performs a local adaptation of the weight-updates according to the behavior of the error function. Contrary to other adaptive techniques, the effect of the RPROP adaptation process is not blurred by the unforeseeable influence of the size of the derivative, but only dependent on the temporal behavior of its sign. This leads t...more

3,170 Citations Source

Cited By353

Newest

Kernel meets recommender systems: A multi-kernel interpolation for matrix completion

Apr 15, 2021·Expert Systems With Applications5.45

#1Zhaoliang Chen (FZU: Fuzhou University)

#2Wei Zhao (CityU: City University of Hong Kong)

Last. Shiping Wang (FZU: Fuzhou University)H-Index: 15

view all 3 authors...

Abstract A primary research direction for recommender systems is matrix completion, which attempts to recover the missing values in a user-item rating matrix. There are numerous approaches for rating tasks, which are mainly classified into latent factor models and neighborhood-based models. Most neighborhood-based models seek similar neighbors by computing similarities in the original data space for final predictions. In this paper, we propose a new neighborhood-based interpolation model with a ...more

Source

Neural graph personalized ranking for Top-N Recommendation

Feb 15, 2021·Knowledge Based Systems5.92

#1Zhibin Hu (SCUT: South China University of Technology)H-Index: 1

#2Jiachun Wang (SCUT: South China University of Technology)

Last. Jin Huang (SCNU: South China Normal University)H-Index: 1

view all 6 authors...

Abstract Personalized recommendation has been widely applied to many real-world services. Many of recent studies focus on collaborative filtering (CF) by deep neural networks, which pursue to predict users’ preference on items based on the past user–item interactions (e.g., a user rates an item). A general CF approach consists of two key modules, embedding representation learning and interaction modeling. In most existing methods, the embedding module is followed by the interaction modeling modu...more

Source

A deep reinforcement learning based long-term recommender system

Feb 15, 2021·Knowledge Based Systems5.92

#1Liwei Huang (University of Electronic Science and Technology of China)H-Index: 1

#2Mingsheng Fu (University of Electronic Science and Technology of China)

Last. Wenyu Chen (University of Electronic Science and Technology of China)H-Index: 6

view all 6 authors...

Abstract Recommender systems aim to maximize the overall accuracy for long-term recommendations. However, most of the existing recommendation models adopt a static view, and ignore the fact that recommendation is a dynamic sequential decision-making process. As a result, they fail to adapt to new situations and suffer from the cold-start problem. Although sequential recommendation methods have been gaining attention recently, the objective of long-term recommendation still has not been explicitl...more

Source

Dig users’ intentions via attention flow network for personalized recommendation

Feb 8, 2021·Information Sciences5.91

#1Yan Chen (SWPU: Southwest Petroleum University)

#2Yongfang Dai (SWPU: Southwest Petroleum University)

Last. Ping Li (SWPU: Southwest Petroleum University)H-Index: 8

view all 6 authors...

Abstract Accurately forecasting user’s purchase intention over time is a huge challenge for personalized recommend systems, of which a critical problem is how to model the changes of user preference and temporal correlation of items. In this paper, aiming at addressing this question, we first introduce attention flow network to model users’ purchase records by leveraging attention flow that describes the changing process of purchase intention. Then based on the attention flow network and individ...more

Source

Learning to utilize auxiliary reviews for recommendation

Feb 4, 2021·Information Sciences5.91

#1Dongmin Hyun (POSTECH: Pohang University of Science and Technology)H-Index: 2

#2Chanyoung Park (UIUC: University of Illinois at Urbana–Champaign)H-Index: 9

Last. Hwanjo Yu (POSTECH: Pohang University of Science and Technology)H-Index: 27

view all 4 authors...

Abstract Review-based recommender systems represent users and items with reviews associated with them. As such, the recommender systems are highly dependent on the number of reviews, which is usually few in number. Thus, they produce inaccurate recommendations to users who rarely wrote reviews. An approach to generate better recommendations for the cold-start users is to augment the scarce reviews using other reviews, which are called auxiliary reviews. In this work, we address two research ques...more

Source

A Hybrid Model- and Memory-Based Collaborative Filtering Algorithm for Baseline Data Prediction of Friedreich's Ataxia Patients

Feb 1, 2021·IEEE Transactions on Industrial Informatics9.11

#1Wenbin YueH-Index: 3

#2Zidong WangH-Index: 113

Last. Xiaohui LiuH-Index: 75

view all 5 authors...

4 Citations Source

Deep learning techniques for rating prediction: a survey of the state-of-the-art

Jan 1, 2021·Artificial Intelligence Review5.75

#1Zahid Younas Khan

#1Zahid Younas Khan (BIT: Beijing Institute of Technology)H-Index: 2

Last. Rukundo PrinceH-Index: 1

view all 4 authors...

With the growth of online information, varying personalization drifts and volatile behaviors of internet users, recommender systems are effective tools for information filtering to overcome the information overload problem. Recommender systems utilize rating prediction approaches i.e. predicting the rating that a user will give to a particular item, to generate ranked lists of items according to the preferences of each user in order to make personalized recommendations. Although previous recomme...more

2 Citations Source

Deep Learning Approach for Matrix Completion Using Manifold Learning.

Dec 11, 2020·arXiv: Learning

#1Saeid Mehrdad

#2Mohammad Hossein Kahaei

Variational auto-encoder based Bayesian Poisson tensor factorization for sparse and imbalanced count data

Dec 10, 2020·Data Mining and Knowledge Discovery2.63

#1Yuan Jin (Monash University)H-Index: 3

#2Ming Liu (Deakin University)H-Index: 3

Last. Yong Xiang (Deakin University)H-Index: 34

view all 7 authors...

Source