FineSID: Scalable and Efficient Semantic Identifier Learning for Generative Recommendation

2026-09-29

AI Summary

This study addresses the challenges of gradient sparsity, low codebook utilization, and severe identifier collisions in generative recommendation, which stem from the Top-1 hard assignment of semantic identifiers. To overcome these issues, this work proposes a unified quantization framework that eliminates reliance on complex initialization priors. By employing differentiable soft assignment, the method distributes learning signals across all codewords, thereby enabling fine-grained gradient propagation and globally balanced codebook optimization. Experimental results demonstrate that the proposed framework significantly improves both codebook utilization and recommendation accuracy. Furthermore, it exhibits strong robustness to initialization configurations, effectively mitigates identifier collisions, and stabilizes the training process.

Abstract

A critical prerequisite of generative recommendation is designing semantic identifiers (SIDs) that are both scalable to large item sets and efficiently learnable. Existing SID learning methods fundamentally rely on Top-1 hard assignment during vector quantization. While heuristic strategies -- such as clustering-based initialization or forced post-hoc collision resolution -- can artificially inflate codebook coverage, they often disrupt end-to-end semantic alignment and fail to address the underlying optimization bottleneck: sparse gradient propagation. In standard Top-1 assignment, gradients concentrate on a narrow subset of frequently selected codewords, leaving the majority inherently under-trained and causing severe SID collisions. To overcome this limitation natively without relying on complex initialization priors, we propose FineSID, a unified quantization framework that moves beyond Top-1 assignment by enabling fine-grained gradient propagation across the entire codebook. Instead of updating only a single selected codeword, FineSID distributes learning signals to all codewords in a soft, differentiable manner. This design promotes globally balanced codebook optimization while strictly preserving semantic consistency, effectively alleviating SID collisions and stabilizing training in large, high-dimensional codebooks. Extensive experiments on multiple public benchmarks demonstrate that FineSID is robust to initialization configurations and consistently improves both codebook utilization and recommendation accuracy. Our work provides a principled, initialization-agnostic solution for semantic identifier learning, advancing the practicality of generative recommendation.

Problem

• Generative Recommendation
• Semantic Identifiers
• Vector Quantization
• Sparse Gradient Propagation
• Codebook Optimization

Innovation

• Semantic Identifier Learning
• Generative Recommendation
• Vector Quantization
• Soft Gradient Propagation
• Codebook Optimization