Bridging the Structural Gap: Adapting Autoregressive Generation for Recommendation

2026-07-23

AI Summary

This work addresses structural degradation and semantic drift in autoregressive generative recommendation, where flattened multi-token item IDs disrupt hierarchical item structures and training–inference codebook mismatches induce representation inconsistencies. To bridge this structural gap, the authors propose three key techniques: Item Context-aware Attention (ICA) to recover structural information during encoding, Hierarchical Path Re-ranking (HPR) and Dual-Path Decoding (DPD) to jointly suppress semantic drift during decoding, and residual quantization for generating semantically meaningful IDs. This is the first approach to systematically reconcile structural modeling and decoding consistency in generative recommendation. Experiments demonstrate state-of-the-art performance on public benchmarks and large-scale offline evaluations; online A/B tests further show significant gains of +0.60% in click-through rate, +1.34% in clicked users, and +1.70% in total reading time.

Abstract

Generative Recommendation (GR) has emerged as a new paradigm for sequential recommendation, in which a representative line of work encodes items into hierarchical semantic IDs via residual quantization and predicts the IDs token by token. However, this generative formulation still exhibits structural gaps with respect to the recommendation task: flattening multi-token IDs into a single sequence destroys item-level structure, and the inconsistency between training and inference over a hierarchical codebook gives rise to semantic drift. To bridge these two gaps, we propose BARGE, which employs Item Context-Aware Attention (ICA) to restore item-level structure during encoding, and Hierarchical Path Reranking (HPR) together with Dual-Path Decoding (DPD) to suppress semantic drift from two complementary angles during decoding. Extensive experiments and analytical studies on public benchmarks and a large-scale offline test demonstrate that BARGE achieves superior recommendation performance. An online A/B test on a Tencent platform yields improvements of 0.60% in click-through rate, 1.34% in click unique visitors, and 1.70% in total reading time, confirming the practical value of BARGE in industrial-scale recommendation.

Problem

• Generative Recommendation
• Structural Gap
• Semantic Drift
• Hierarchical Codebook
• Item-level Structure

Innovation

• Generative Recommendation
• Item Context-Aware Attention
• Hierarchical Path Reranking
• Dual-Path Decoding
• Semantic Drift