Multi-Behavior Generative Recommendation

TL;DR

This paper introduces MBGen, a generative framework for multi-behavior sequential recommendation that models user behavior and item prediction as a unified autoregressive process, improving recommendation accuracy.

Contribution

MBGen formulates multi-behavior recommendation as a two-step generative process and proposes a position-routed sparse architecture for efficient modeling.

Findings

MBGen significantly outperforms existing models on public datasets.

The unified generative approach effectively captures behavior-item interactions.

The position-routed sparse architecture scales well to large models.

Abstract

Multi-behavior sequential recommendation (MBSR) aims to incorporate behavior types of interactions for better recommendations. Existing approaches focus on the next-item prediction objective, neglecting the value of integrating the target behavior type into the learning objective. In this paper, we propose MBGen, a novel Multi-Behavior sequential Generative recommendation framework. We formulate the MBSR task into a consecutive two-step process: (1) given item sequences, MBGen first predicts the next behavior type to frame the user intention, (2) given item sequences and a target behavior type, MBGen then predicts the next items. To model such a two-step process, we tokenize both behaviors and items into tokens and construct one single token sequence with both behaviors and items placed interleaved. Furthermore, MBGen learns to autoregressively generate the next behavior and item tokens in a unified generative recommendation paradigm, naturally enabling a multi-task capability. Additionally, we exploit the heterogeneous nature of token sequences in the generative recommendation and propose a position-routed sparse architecture to efficiently and effectively scale up models. Extensive experiments on public datasets demonstrate that MBGen significantly outperforms existing MBSR models across multiple tasks.