Learning representations using Deep learning for Recommendation

The objective of this class is to present deep representation learning methods for recommendation task. This class was presented at SFdS.

Contents

Introduction to deep learning

Multi-layer neural network

• Practical exercise on building deep network for recommendation

Learning user and item representations

• Learning item representation using Convolution Neural Network
• Practical exercise on adding movie poster representation into deep network
• Learning user representation using Recurrent Neural Network

The practical exercise shows how to build deep network for recommendation on MovieLens 1M dataset.

Practical details.

AWS EC2 public image: ami-49cff25f is configured to run the exercise.