End-to-End Semantic ID Generation for Generative Advertisement Recommendation

2026-02-11

AI Summary

Existing semantic ID (SID) generation methods in generative recommendation rely on two-stage residual quantization, which suffers from objective misalignment, semantic degradation, and error accumulation. This work proposes UniSID, a novel framework that achieves the first end-to-end unified optimization for SID generation. By employing multi-granularity contrastive learning to align semantic representations across different levels and introducing a summarization-based ad reconstruction mechanism to enhance high-level semantic expression, UniSID significantly improves SID quality while preserving the generative recommendation architecture. Experimental results demonstrate that the proposed method outperforms state-of-the-art SID generation approaches, achieving up to a 4.62% relative improvement in Hit Rate on downstream ad recommendation tasks.

Technology Category

• Natural Language Processing: Generation
• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Search and Optimization: Learning to Search

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Generative Recommendation (GR) has excelled by framing recommendation as next-token prediction. This paradigm relies on Semantic IDs (SIDs) to tokenize large-scale items into discrete sequences. Existing GR approaches predominantly generate SIDs via Residual Quantization (RQ), where items are encoded into embeddings and then quantized to discrete SIDs. However, this paradigm suffers from inherent limitations: 1) Objective misalignment and semantic degradation stemming from the two-stage compression; 2) Error accumulation inherent in the structure of RQ. To address these limitations, we propose UniSID, a Unified SID generation framework for generative advertisement recommendation. Specifically, we jointly optimize embeddings and SIDs in an end-to-end manner from raw advertising data, enabling semantic information to flow directly into the SID space and thus addressing the inherent limitations of the two-stage cascading compression paradigm. To capture fine-grained semantics, a multi-granularity contrastive learning strategy is introduced to align distinct items across SID levels. Finally, a summary-based ad reconstruction mechanism is proposed to encourage SIDs to capture high-level semantic information that is not explicitly present in advertising contexts. Experiments demonstrate that UniSID consistently outperforms state-of-the-art SID generation methods, yielding up to a 4.62% improvement in Hit Rate metrics across downstream advertising scenarios compared to the strongest baseline.

Problem

• Semantic ID
• Generative Recommendation
• Residual Quantization
• Objective Misalignment
• Error Accumulation

Innovation

• End-to-End Semantic ID
• Generative Recommendation
• Multi-Granularity Contrastive Learning
• Ad Reconstruction
• Unified SID Generation