Aug 2026· Proceedings of the 32nd ACM SIGKDD Conference on Knowledge Discovery and Data Mining V.2· pp. 1274-1285· 0 citations· 35 references
Abstract
Generative recommendation (GR) reformulates sequential recommendation as an autoregressive generation problem, where items are represented as discrete semantic IDs. However, learning effective item tokenization is critical yet remains challenging. Most existing methods optimize tokenization in a point-wise or heuristic manner, which often leads to Relational Structure Distortion. We identify two main sources of this distortion: (1) point-wise reconstruction losses that are insufficient to preserve item--item relations, and (2) collision-mitigation--driven code assignment that disrupts semantic neighborhoods under many-to-one mappings. This issue is especially harmful for recommendation, since the task fundamentally depends on relative item similarity. To address this issue, we propose ReSOT, a unified framework that Re-balances Semantic ID learning via Optimal Transport for GR. ReSOT consists of two complementary components: a structure-aware reconstruction module that aligns item–item relations between the original and the reconstructed embedding space, and an unbalanced optimal transport-based soft quantization module that achieves a better trade-off between collision mitigation and semantic neighborhood preservation. Extensive experiments on real-world datasets show that ReSOT provides a principled tokenization scheme that preserves relational structure while assigning codes in a collision-aware and semantics-consistent manner. Our code is available at https://github.com/grw-zju/ReSOT.
BARGE is proposed, 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.
Difficulty-Aware Semantic-ID Optimization (DASO), a tree-aware post-training method that addresses failure mode as an online rollout-allocation problem and improves over MiniOneRec-style GRPO on 11 of 12 metrics and achieves the best result on 9 of 12 metrics.
This work proposes HiST (Hierarchical Semantic Tree Augmentation), a novel framework that leverages branch information to bridge the semantic gap and can be effectively integrated with various existing tokenization methods to enhance recommendation model performance.
Bocheng Pan, Hai-Long Shi, Xingyu Gao· Proceedings of the 32nd ACM...· 0 citations
This work proposes a simple, parameter-free intervention that initializes SID token embeddings directly from their corresponding centroids in the semantic embedding space, and shows that preserving SID geometry, beyond shared-prefix structure, provides a simple and effective semantic prior for LLM-based GR.
Donald Loveland, Liam Collins, B. Kumar et al.· 0 citations
Large Language Models (LLMs) have demonstrated remarkable general-purpose abilities across a wide range of domains, and these strengths have also been increasingly evidenced in recommender systems. However, existing methods that attempt to integrate collaborative signals into LLMs often fail to preserve their foundational knowledge. This loss is critical in text-rich recommendation, where robust semantic understanding is required to interpret user reviews and item profiles. We propose PALRec, a parameter-preserving augmentation framework that equips an LLM with recommendation capabilities while keeping its original parameters fixed. We first construct evidence-grounded user and item profiles from reviews and use them as concise pseudo-labels for reconstruction. We then introduce lightweight, trainable user and item embedding modules optimized with a multi-task objective that combines next-item prediction and profile reconstruction. These modules are trained jointly to align collaborative signals with the LLM’s semantic space without modifying the backbone. We also employ token-aware loss decomposition and frequency-aware reweighting to stabilize training and mitigate popularity bias. Experiments on public benchmarks show that PALRec consistently outperforms fully fine-tuned counterparts in recommendation accuracy while preserving the LLM’s pre-trained knowledge. This result highlights that maintaining the LLM’s semantic understanding is crucial for effectively exploiting textual information in recommender systems.
Hyunsoo Na, Minseok Gang, Sang-goo Lee et al.· ACM Transactions on Informat...· 0 citations
Cross-domain sequential recommendation (CDSR) improves target-domain prediction by leveraging multi-domain interaction histories. Most CDSR methods rely on shared entities or co-occurrence signals, which become unreliable when overlap is limited, and atomic ID representations further generalize poorly to long-tail or unseen items as cross-domain distribution shifts exacerbate this problem. Recent generative CDSR methods enable cross-domain transfer without relying on raw ID alignment by generating content-grounded semantic IDs (SIDs) for cross-domain alignment. However, two challenges remain, including (i) user-item asymmetry, with items discretized for generation whereas user preferences are encoded only implicitly in sequence representations, limiting semantic-level preference control; and (ii) selective transfer, making it difficult to assess source-domain signals against the target preference representation without an explicit discrete user anchor aligned with item IDs, which can lead to unintended transfer of irrelevant signals. This work presents UniGCRec, which constructs a cross-domain user profile from multi-domain histories and quantizes both users and items into CSC-IDs that integrate semantic and collaborative signals. This symmetric quantization places user and item representations in the same discrete CSC-ID space, effectively mitigating user-item asymmetry and enabling preference-aware selective transfer under low-overlap settings. The generator is conditioned on a user CSC-ID prefix and the target domain item CSC-ID history for next-item generation, with trie-constrained decoding ensuring target domain validity. Experiments on public multi-domain benchmarks show consistent gains over strong baselines, with particularly strong gains on several target domains.
Chaoyue Ding, Jia-Hao Liu, Dongsheng Li et al.· Proceedings of the 32nd ACM...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
