IntRR: A Framework for Integrating SID Redistribution and Length Reduction

TL;DR

IntRR introduces a novel framework that aligns SID objectives with recommendation goals and reduces sequence length, significantly improving efficiency and accuracy in generative recommendation systems.

Contribution

It proposes a new method that dynamically redistributes semantic weights and handles SID hierarchies recursively, addressing sequence inflation and static identifiers issues.

Findings

Outperforms baseline models in recommendation accuracy

Reduces inference latency and computational overhead

Achieves superior efficiency without sacrificing performance

Abstract

Generative Recommendation (GR) has emerged as a transformative paradigm that reformulates the traditional cascade ranking system into a sequence-to-item generation task, facilitated by the use of discrete Semantic IDs (SIDs). However, current SIDs are suboptimal as the indexing objectives (Stage 1) are misaligned with the actual recommendation goals (Stage 2). Since these identifiers remain static (Stage 2), the backbone model lacks the flexibility to adapt them to the evolving complexities of user interactions. Furthermore, the prevailing strategy of flattening hierarchical SIDs into token sequences leads to sequence length inflation, resulting in prohibitive computational overhead and inference latency. To address these challenges, we propose IntRR, a novel framework that integrates objective-aligned SID Redistribution and structural Length Reduction. By leveraging item-specific Unique IDs (UIDs) as collaborative anchors, this approach dynamically redistributes semantic weights across hierarchical codebook layers. Concurrently, IntRR handles the SID hierarchy recursively, eliminating the need to flatten sequences. This ensures a fixed cost of one token per item. Extensive experiments on benchmark datasets demonstrate that IntRR yields substantial improvements over representative generative baselines, achieving superior performance in both recommendation accuracy and efficiency.