Semantic Trimming and Auxiliary Multi-step Prediction for Generative Recommendation

[2604.05329] Semantic Trimming and Auxiliary Multi-step Prediction for Generative Recommendation Skip to main content Search Submit Donate Log in Computer Science > Information Retrieval arXiv:2604.05329 (cs) [Submitted on 7 Apr 2026 ( v1 ), last revised 28 Aug 2026 (this version, v3)] Title: Semantic Trimming and Auxiliary Multi-step Prediction for Generative Recommendation Authors: Tianyu Zhan , Kairui Fu , Chengfei Lv , Zheqi Lv , Shengyu Zhang View PDF HTML (experimental) Abstract: Generative Recommendation (GR) has recently transitioned from atomic item-indexing to Semantic ID (SID)-based frameworks to capture intrinsic item relationships and enhance generalization. However, the adoption of high-granularity SIDs leads to two critical challenges: prohibitive training overhead due to sequence expansion and unstable performance reliability characterized by non-monotonic accuracy fluctuations. We identify that these disparate issues are fundamentally rooted in the Semantic Dilution Effect, where redundant tokens waste massive computation and dilute the already sparse learning signals in recommendation. To counteract this, we propose STAMP (Semantic Trimming and Auxiliary Multi-step Prediction), a framework utilizing a dual-end optimization strategy. We argue that effective SID learning requires simultaneously addressing low input information density and sparse output supervision. On the input side, Semantic Adaptive Pruning (SAP) dynamically filters redundancy during the forward pass, converting noise-laden sequences into compact, information-rich representations. On the output side, Multi-step Auxiliary Prediction (MAP) employs a multi-token objective to densify feedback, strengthening long-range dependency capture and ensuring robust learning signals despite compressed inputs. Unifying input purification and signal amplification, STAMP enhances both training efficiency and representation capability. Experiments on public Amazon and large-scale industrial datasets show STAMP achieves 1.23--1.38$\times$ speedup and 17.2\%--54.7\% VRAM reduction while maintaining or improving performance across multiple architectures. Comments: 9 pages, Under Review Subjects: Information Retrieval (cs.IR) Cite as: arXiv:2604.05329 [cs.IR] (or arXiv:2604.05329v3 [cs.IR] for this version) https://doi.org/10.48550/arXiv.2604.05329 Focus to learn more arXiv-issued DOI via DataCite Submission history From: Tianyu Zhan [ view email ] [v1] Tue, 7 Apr 2026 02:00:23 UTC (3,156 KB) [v2] Wed, 19 Aug 2026 03:37:02 UTC (4,807 KB) [v3] Fri, 28 Aug 2026 07:38:54 UTC (4,808 KB) Full-text links: Access Paper: View PDF HTML (experimental) TeX Source view license Additional Features Audio Summary Current browse context: cs.IR < prev | next > new | recent | 2026-04 Change to browse by: cs References & Citations SciX Google Scholar Semantic Scholar export BibTeX citation Bookmark Bibliographic Tools Bibliographic and Citation Tools Bibliographic Explorer Toggle Bibliographic Explorer ( What is the Explorer? ) Connected Papers Toggle Connected Papers ( What is Connected Papers? ) Litmaps Toggle Litmaps ( What is Litmaps? ) scite.ai Toggle scite Smart Citations ( What are Smart Citations? ) Code, Data, Media Code, Data and Media Associated with this Article alphaXiv Toggle alphaXiv ( What is alphaXiv? ) Links to Code Toggle CatalyzeX Code Finder for Papers ( What is CatalyzeX? ) DagsHub Toggle DagsHub ( What is DagsHub? ) GotitPub Toggle Gotit.pub ( What is GotitPub? ) Huggingface Toggle Hugging Face ( What is Huggingface? ) ScienceCast Toggle ScienceCast ( What is ScienceCast? ) Demos Demos Replicate Toggle Replicate ( What is Replicate? ) Spaces Toggle Hugging Face Spaces ( What is Spaces? ) Spaces Toggle TXYZ.AI ( What is TXYZ.AI? ) Related Papers Recommenders and Search Tools Link to Influence Flower Influence Flower ( What are Influence Flowers? ) Core recommender toggle CORE Recommender ( What is CORE? ) Author Venue Institution Topic About arXivLabs arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs . Which authors of this paper are endorsers? | Disable MathJax ( What is MathJax? ) We gratefully acknowledge support from our major funders , member institutions , and all contributors. About · Help · Contact · Subscribe · Copyright · Privacy · Accessibility · Operational Status (opens in new tab)

Related documents

• Semantic Trimming and Auxiliary Multi-step Prediction for Generative Recommendation#682636
• Modeling Behavioral Intensity and Transitions for Generative Recommendation#138941
• Bumblebee: Interleaved Mixed-Layer Building Blocks for Large-Scale Recommendation Systems#1011172
• TurboGR: An Accelerated Training System for Large-Scale Generative Recommendation#180626
• CoFiRec: Coarse-to-Fine Tokenization for Generative Recommendation#798680