Harmonizing Generative Retrieval and Ranking in Chain-of-Recommendation

TL;DR

RecoChain is a unified Transformer-based framework that combines generative retrieval and ranking to improve recommendation quality by integrating candidate generation with click probability estimation.

Contribution

It introduces a novel unified framework that bridges the gap between generative retrieval and ranking in recommender systems using a single Transformer model.

Findings

Achieves improved Top-K recommendation performance.

Effectively bridges the gap between generative retrieval and ranking.

Maintains strong generative capability.

Abstract

Generative recommender systems have recently emerged as a promising paradigm by formulating next-item prediction as an auto-regressive semantic IDs generation, such as OneRec series works. However, with the next-item-agnostic prediction paradigm, its could beam out some next potential items via Semantic IDs but hard to estimate which items are better from them, e.g., select the top-10 from beam-256 items, leading to a gap between generation and ranking performance. To fulfill this gap, we propose RecoChain, a unified generative retrieval and ranking framework that integrates candidate generation and ranking within a single Transformer backbone. Specifically, in inference, the model first generates candidate items via hierarchical semantic ID prediction, then performs the SIM-based ranking process to estimate the click possibility of corresponding item candidate continuously. Extensive experiments on large-scale real-world datasets demonstrate that our approach effectively bridges the gap between generative retrieval and ranking, achieving improved Top-K recommendation performance while maintaining strong generative capability.