Description
Wide&Deep jointly trains wide linear models and deep neural networks to combine the benefits of memorization and generalization for real-world recommender systems. In summary, the wide component is a generalized linear model. The deep component is a feed-forward neural network. The deep and wide components are combined using a weighted sum of their output log odds as the prediction. This is then fed to a logistic loss function for joint training, which is done by back-propagating the gradients from the output to both the wide and deep part of the model simultaneously using mini-batch stochastic optimization. The AdaGrad optimizer is used for the wider part. The combined model is illustrated in the figure (center).
Papers Using This Method
Knowledge Graph Driven Recommendation System Algorithm2023-12-01TFNet: Multi-Semantic Feature Interaction for CTR Prediction2020-06-29Click-Through Rate Prediction with the User Memory Network2019-07-09Structured Semantic Model supported Deep Neural Network for Click-Through Rate Prediction2018-12-04Neural Factorization Machines for Sparse Predictive Analytics2017-08-16Attentional Factorization Machines: Learning the Weight of Feature Interactions via Attention Networks2017-08-15Wide & Deep Learning for Recommender Systems2016-06-24