Latent User Intent Modeling for Sequential Recommenders

TL;DR

This paper introduces a probabilistic approach using variational autoencoders to model latent user intents in sequential recommender systems, enhancing understanding of user behavior for improved recommendations.

Contribution

It presents a novel latent intent modeling framework that integrates user intent inference into sequential recommendation, which was not previously addressed.

Findings

Improved recommendation accuracy in offline tests.

Enhanced user engagement in live experiments.

Effective inference of user intents from behavior signals.

Abstract

Sequential recommender models are essential components of modern industrial recommender systems. These models learn to predict the next items a user is likely to interact with based on his/her interaction history on the platform. Most sequential recommenders however lack a higher-level understanding of user intents, which often drive user behaviors online. Intent modeling is thus critical for understanding users and optimizing long-term user experience. We propose a probabilistic modeling approach and formulate user intent as latent variables, which are inferred based on user behavior signals using variational autoencoders (VAE). The recommendation policy is then adjusted accordingly given the inferred user intent. We demonstrate the effectiveness of the latent user intent modeling via offline analyses as well as live experiments on a large-scale industrial recommendation platform.