Experiments on three public sequential-recommendation benchmarks show that SID Routing improves its corresponding direct-generation, and indicate that structure-grounded reasoning is most useful when prefix matching is insufficient but learnable SID-space transitions remain available, whereas long or weakly supported routes introduce additional decoding cost and accumulated errors.
Abstract
Semantic-ID-based generative recommendation represents each item as a hierarchical discrete token sequence and reformulates next-item prediction as constrained sequence generation. Existing methods, however, mainly use Semantic IDs as target sequences to be memorized, leaving the hierarchy, intra-layer relations, and item neighborhoods underused as an explicit reasoning space. Explicit reasoning-enhanced generative methods often produce a natural-language rationale before the item identifier, but this rationale is only weakly coupled with the discrete SID space in which the final prediction is made. We propose CogRec, a structure-cognitive fast-and-slow reasoning framework that grounds intermediate reasoning in the same SID topology used for target generation. CogRec augments the vertical SID hierarchy with intra-layer semantic graphs and item-level neighborhoods, and introduces SID Routing to represent recommendation reasoning through layer-wise Match, LateralJump, and Explore operations. Exact matching implements fast semantic localization, whereas lateral and exploratory operations instantiate slower structural navigation. A supervised multi-stage pipeline aligns the newly introduced SID tokens, establishes direct SID generation, and trains natural-language and SID-routing reasoning branches from a shared checkpoint under the same trie-constrained output space. Experiments on three public sequential-recommendation benchmarks show that SID Routing improves its corresponding direct-generation, indicate that structure-grounded reasoning is most useful when prefix matching is insufficient but learnable SID-space transitions remain available, whereas long or weakly supported routes introduce additional decoding cost and accumulated errors. Code is available at https://github.com/caskcsg/CogRec
Semantic ID-based generative recommendation predicts an item by generating a short sequence of semantic ID tokens, where each token is produced autoregressively. Latent reasoning has recently been introduced to improve this process through additional hidden-state computation before each token decision. This raises a practical question: when one item is represented by a sequence of semantic ID tokens, should each token receive the same fixed number of latent refinement steps, or should these steps be allocated more effectively across positions? We study this question through position-wise information-gain (IG), which measures how much each semantic ID position reduces the uncertainty of the target item. We observe that earlier semantic ID positions usually provide higher information-gain, while later positions contribute less additional information. We further analyze that applying more refinement to high-IG positions tends to bring larger expected benefits. Based on this observation, we propose IBA, an Information-Gain Budget Allocation framework for semantic ID-based generative recommendation. IBA treats latent refinement steps as a limited computational resource and learns how to allocate them across semantic ID positions, assigning more refinement to informative positions and less to positions with smaller contribution. Experiments on multiple public datasets show that IBA consistently improves strong generative recommendation baselines and achieves a better accuracy--computation trade-off than fixed or poorly matched step allocations.
Shangxin Yang, Min Gao, Zongwei Wang et al.· 0 citations
Generative recommendation (GR) assigns each item content-derived Semantic IDs (SIDs) and employs a sequence-to-sequence model to generate the target item's SIDs based on past user sessions. This paradigm typically outperforms sequential recommendation (SR) methods that rely solely on behavior modeling. However, GR methods suffer from semantic preference bias, where they favor semantically similar items while neglecting items with weak semantic similarity but manifest behavioral relevance--patterns that SR models capture more effectively through behavior modeling. In this work, we propose OMG, a GR method that unifies behavior modeling and semantic generation to produce recommendations that seamlessly integrate both information. First, we introduce Hybrid Behavior-Semantic Encoding, enabling OMG to capture salient user behavioral patterns in addition to semantic correlation when encoding interaction sequences, aiming to recall items with certain behavioral relevance but weak semantic similarity. Then, we propose Behavior-Semantic Latent Reasoning. It iteratively refines the encoder output, progressively transforming it from the behavioral space to the semantic space, and leverages the resulting trajectory to guide SIDs decoding, thereby bridging behavior modeling and semantic generation. Finally, we employ Behavior-Reinforced Semantic Generation, which introduces reinforcement learning with behavior-aware rewards to enhance the model's reasoning capability and better align the decoder's generation process with user behavior. Experiments on benchmark datasets demonstrate that OMG consistently outperforms baseline methods.
Binquan Wu, Xinbo Chen, Yicheng Luo et al.· Proceedings of the 32nd ACM...· 0 citations
Semantic-ID-based generative recommendation represents items as sequences of shared semantic tokens, enabling token recombination beyond isolated item IDs. However, closed-world recombination does not necessarily imply temporal open-token cold-start induction, where new items enter the item catalog with unseen atomic tokens or weakly supported SID paths. In this work, we revisit SID-based generative recommendation under an absolute-time temporal protocol that separates seen and unseen targets and diagnoses the cold item reachability at the token level. Through seen/unseen-hit analysis, coldness taxonomy, and oracle-prefix probing, we show that current SID-based models can occasionally reach future items supported by observed tokens and prefixes, but struggle with unseen atomic tokens and unsupported SID paths. We further explain this boundary by interpreting SID generation as hierarchical semantic bucketing: early tokens select coarse semantic regions, while later tokens refine item-specific paths. These findings show that SID generation is compositional but not fully open-ended, and suggest future directions in more independent SID spaces, scoring-based interfaces, and dynamic textual context.
Jie Peng, Yanping Zheng, Zhewei Zhe et al.· 1 citation
This work proposes GUIDER (Generative User Interest Discovery & Explicit Reasoning), a framework that fundamentally decouples intent planning from item matching by reformulating sequential modeling within a rigorous closed-set semantic interest space.
Jinke Wu, Yinghao Wu, Shuchang Liu et al.· Proceedings of the 32nd ACM...· 0 citations
Traditional sequential recommender systems have primarily relied on implicit pattern recognition in user interaction sequences, achieving strong performance but functioning as “black boxes” that lack transparent reasoning. This paper introduces LLM-RS, a novel framework that leverages Large Language Models to enable explicit reasoning chains in sequential recommendation. Our approach transforms the recommendation task from mere pattern matching to interpretable reasoning by developing a multi-stage architecture that: (1) extracts structured preference profiles from user interaction sequences, (2) generates explicit reasoning chains analyzing candidate items against inferred preferences, and (3) produces persuasive explanations alongside recommendations. We propose three model variants—fine-tuned reasoning, retrieval-augmented generation, and hybrid ensemble—that integrate LLM-based reasoning with traditional collaborative filtering. The framework addresses key challenges in modern recommender systems by providing transparent, persuasive rationales while maintaining competitive performance, marking a significant step toward more interpretable and trustworthy recommendation systems. Comprehensive evaluations across the Amazon Reviews, MovieLens, MIND, and KuaiSAR datasets demonstrate that LLM-RS not only matches state-of-the-art methods in accuracy but also significantly enhances explanation quality, user trust, and recommendation diversity. Our findings reveal that reasoning-enabled recommendations increase user adherence in online experiments and improve long-term engagement metrics.
A. Gadallah, H. Hefny, Mohammed E. Almandouh et al.· Mathematics· 0 citations
Industrial recommender systems typically rely on cascaded retrieval, pre-ranking, ranking, and reranking stages, whose separately optimized models limit scaling, fragment decision making, and lack semantic knowledge and reasoning. We present TGR (Tencent Generative Recommendation), an industrial framework that advances recommendation toward the generative paradigm along three coupled directions. TGR-GenRank upgrades ranking through CCFormer, which combines unified feature tokenization, a scalable Transformer backbone, feature-field separated cross attention, subspace token mixing, and hierarchical sequence compression while retaining per-item multi-task outputs. TGR-GenRec explores end-to-end generation under two paradigms: BARGE bridges item-boundary loss and semantic drift in hierarchical semantic-ID generation through item context-aware attention, hierarchical path reranking, and orthogonal dual-path decoding; HiGR performs whole-slate generation with prefix-structured semantic IDs, coarse-to-fine decoding, and listwise multi-objective alignment. TGR-Reason injects offline-generated semantic-ID reason tokens into online decoding, providing reasoning without request-time rollout. TGR is deployed across Tencent production surfaces serving hundreds of millions of users. CCFormer delivers significant gains in five A/B-tested scenarios and is fully launched in two, including +3.57% CTR and +1.71% advertising revenue. BARGE improves Hit@5 by 10.2-16.9% and yields +0.60% CTR and +1.70% reading time after full rollout. HiGR improves offline slate quality by 15.9-21.3% with a 5x inference speedup and achieves up to +1.22% watch time and +1.73% video views. TGR-Reason raises cold-start new-user Hit@1 by 477.8% and delivers +1.75% effective consumption and +13.09% new-user exposure-to-conversion online.
Tgr Team Lei Cheng, Hao-Nan Hu, Beibei Kong et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
