Metric Learning for Session-Based Recommendations

Szymon Zaborowski

A session-based recommendation algorithm with the usage of metric learning is an approach that can outperform existing methods simultaneously being less complex than other approaches like GRU4REC. Here we will show how our algorithm works, why metric learning is a good direction for recommendations tasks, and how to implement this solution to session recommendations using popular deep learning libraries like TensorFlow or PyTorch.