Multi-Interest Network with Dynamic Routing for Recommendation at Tmall

TL;DR

This paper introduces MIND, a multi-interest recommendation model using dynamic routing to better capture diverse user interests, significantly improving recommendation accuracy and deployed at Tmall's mobile app.

Contribution

The paper proposes a novel multi-interest network with capsule routing and label-aware attention, advancing user interest modeling for large-scale industrial recommender systems.

Findings

MIND outperforms state-of-the-art methods on public benchmarks.

MIND achieves superior recommendation accuracy in large-scale industrial data.

Deployed at Tmall, MIND handles major online traffic effectively.

Abstract

Industrial recommender systems usually consist of the matching stage and the ranking stage, in order to handle the billion-scale of users and items. The matching stage retrieves candidate items relevant to user interests, while the ranking stage sorts candidate items by user interests. Thus, the most critical ability is to model and represent user interests for either stage. Most of the existing deep learning-based models represent one user as a single vector which is insufficient to capture the varying nature of user's interests. In this paper, we approach this problem from a different view, to represent one user with multiple vectors encoding the different aspects of the user's interests. We propose the Multi-Interest Network with Dynamic routing (MIND) for dealing with user's diverse interests in the matching stage. Specifically, we design a multi-interest extractor layer based on capsule routing mechanism, which is applicable for clustering historical behaviors and extracting diverse interests. Furthermore, we develop a technique named label-aware attention to help learn a user representation with multiple vectors. Through extensive experiments on several public benchmarks and one large-scale industrial dataset from Tmall, we demonstrate that MIND can achieve superior performance than state-of-the-art methods for recommendation. Currently, MIND has been deployed for handling major online traffic at the homepage on Mobile Tmall App.