Heng-Tze Cheng
ABSTRACT
Generalized linear models with nonlinear feature transformations are widely used for large-scale regression and classification problems with sparse inputs. Memorization of feature interactions through a wide set of cross-product feature transformations are effective and interpretable, while generalization requires more feature engineering effort. With less feature engineering, deep neural networks can generalize better to unseen feature combinations through low-dimensional dense embeddings learned for the sparse features. However, deep neural networks with embeddings can over-generalize and recommend less relevant items when the user-item interactions are sparse and high-rank. In this paper, we present Wide & Deep learning---jointly trained wide linear models and deep neural networks---to combine the benefits of memorization and generalization for recommender systems. We productionized and evaluated the system on Google Play, a commercial mobile app store with over one billion active users and over one million apps. Online experiment results show that Wide & Deep significantly increased app acquisitions compared with wide-only and deep-only models. We have also open-sourced our implementation in TensorFlow.
- J. Duchi, E. Hazan, and Y. Singer. Adaptive subgradient methods for online learning and stochastic optimization. Journal of Machine Learning Research, 12:2121--2159, July 2011. Google Scholar
Digital Library
- K. He, X. Zhang, S. Ren, and J. Sun. Deep residual learning for image recognition. Proc. IEEE Conference on Computer Vision and Pattern Recognition, 2016.Google ScholarCross Ref
- H. B. McMahan. Follow-the-regularized-leader and mirror descent: Equivalence theorems and l1 regularization. In Proc. AISTATS, 2011.Google Scholar
- T. Mikolov, A. Deoras, D. Povey, L. Burget, and J. H. Cernocky. Strategies for training large scale neural network language models. In IEEE Automatic Speech Recognition & Understanding Workshop, 2011.Google ScholarCross Ref
- S. Rendle. Factorization machines with libFM. ACM Trans. Intell. Syst. Technol., 3(3):57:1--57:22, May 2012. Google ScholarDigital Library
- J. J. Tompson, A. Jain, Y. LeCun, and C. Bregler. Joint training of a convolutional network and a graphical model for human pose estimation. In Z. Ghahramani, M. Welling, C. Cortes, N. D. Lawrence, and K. Q. Weinberger, editors, NIPS, pages 1799--1807. 2014. Google ScholarDigital Library
- H. Wang, N. Wang, and D.-Y. Yeung. Collaborative deep learning for recommender systems. In Proc. KDD, pages 1235--1244, 2015. Google ScholarDigital Library
- B. Yan and G. Chen. AppJoy: Personalized mobile application discovery. In MobiSys, pages 113--126, 2011. Google ScholarDigital Library
- Wide & Deep Learning for Recommender Systems
Recommendations
Wide & Deep Learning in Job Recommendation: An Empirical Study
Information Retrieval TechnologyAbstractRecommender systems have become more and more popular in recent years. Collaborative Filtering and Content-Based methods are widely used for a long time. Recently, some researchers introduced deep learning algorithms into recommender system. In ...
Acquiring User Information Needs for Recommender Systems
WI-IAT '13: Proceedings of the 2013 IEEE/WIC/ACM International Joint Conferences on Web Intelligence (WI) and Intelligent Agent Technologies (IAT) - Volume 03Most recommender systems attempt to use collaborative filtering, content-based filtering or hybrid approach to recommend items to new users. Collaborative filtering recommends items to new users based on their similar neighbours, and content-based ...
A Deep Autoencoder-Based Hybrid Recommender System
Recommender systems build their suggestions on rating data, given explicitly or implicitly by users on items. These ratings create a huge sparse user-item rating matrix which opens two challenges for researchers on the field. The first challenge is ...
Comments