DVE: Dynamic Variational Embeddings with Applications in Recommender Systems

TL;DR

This paper introduces DVE, a dynamic variational embedding method leveraging recurrent neural networks to model temporal variations in sequence-aware data, significantly improving recommender system performance.

Contribution

The paper presents a novel dynamic embedding approach that explicitly models temporal variations and intrinsic features, advancing sequence-aware recommendation techniques.

Findings

DVE outperforms static embedding methods in link prediction tasks.

The approach effectively captures temporal dynamics in node features.

End-to-end neural architecture enhances recommendation accuracy.

Abstract

Embedding is a useful technique to project a high-dimensional feature into a low-dimensional space, and it has many successful applications including link prediction, node classification and natural language processing. Current approaches mainly focus on static data, which usually lead to unsatisfactory performance in applications involving large changes over time. How to dynamically characterize the variation of the embedded features is still largely unexplored. In this paper, we introduce a dynamic variational embedding (DVE) approach for sequence-aware data based on recent advances in recurrent neural networks. DVE can model the node's intrinsic nature and temporal variation explicitly and simultaneously, which are crucial for exploration. We further apply DVE to sequence-aware recommender systems, and develop an end-to-end neural architecture for link prediction.