惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

博客园 - 三生石上(FineUI控件)
freeCodeCamp Programming Tutorials: Python, JavaScript, Git & More
月光博客
月光博客
博客园 - 【当耐特】
Hugging Face - Blog
Hugging Face - Blog
让小产品的独立变现更简单 - ezindie.com
让小产品的独立变现更简单 - ezindie.com
爱范儿
爱范儿
大猫的无限游戏
大猫的无限游戏
S
SegmentFault 最新的问题
博客园_首页
雷峰网
雷峰网
量子位
有赞技术团队
有赞技术团队
博客园 - 叶小钗
博客园 - 聂微东
V
V2EX
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
博客园 - 司徒正美
小众软件
小众软件
The Cloudflare Blog
阮一峰的网络日志
阮一峰的网络日志
Apple Machine Learning Research
Apple Machine Learning Research
Jina AI
Jina AI
人人都是产品经理
人人都是产品经理

cs.IR updates on arXiv.org

From Top-1 to Top-K: A Reproducibility Study and Benchmarking of Counterfactual Explanations for Recommender Systems Impact of large language models on peer review opinions from a fine-grained perspective: Evidence from top conference proceedings in AI Diagnosable ColBERT: Debugging Late-Interaction Retrieval Models Using a Learned Latent Space as Reference Enhancing Unsupervised Keyword Extraction in Academic Papers through Integrating Highlights with Abstract CAST: Modeling Semantic-Level Transitions for Complementary-Aware Sequential Recommendation IndiaFinBench: An Evaluation Benchmark for Large Language Model Performance on Indian Financial Regulatory Text Think Before Writing: Feature-Level Multi-Objective Optimization for Generative Citation Visibility RARE: Redundancy-Aware Retrieval Evaluation Framework for High-Similarity Corpora Personalized Benchmarking: Evaluating LLMs by Individual Preferences Modular Representation Compression: Adapting LLMs for Efficient and Effective Recommendations JFinTEB: Japanese Financial Text Embedding Benchmark UsefulBench: Towards Decision-Useful Information as a Target for Information Retrieval SIMMER: Cross-Modal Food Image--Recipe Retrieval via MLLM-Based Embedding Rethinking the Necessity of Adaptive Retrieval-Augmented Generation through the Lens of Adaptive Listwise Ranking BioHiCL: Hierarchical Multi-Label Contrastive Learning for Biomedical Retrieval with MeSH Labels Learning Behaviorally Grounded Item Embeddings via Personalized Temporal Contexts Collaborative Filtering Through Weighted Similarities of User and Item Embeddings IG-Search: Step-Level Information Gain Rewards for Search-Augmented Reasoning Metric-agnostic Learning-to-Rank via Boosting and Rank Approximation GenRec: A Preference-Oriented Generative Framework for Large-Scale Recommendation Uncertainty-aware Generative Learning Path Recommendation with Cognition-Adaptive Diffusion CPGRec+: A Balance-oriented Framework for Personalized Video Game Recommendations Don't Retrieve, Navigate: Distilling Enterprise Knowledge into Navigable Agent Skills for QA and RAG NewsTorch: A PyTorch-based Toolkit for Learner-oriented News Recommendation Controlling Authority Retrieval: A Missing Retrieval Objective for Authority-Governed Knowledge APEX-MEM: Agentic Semi-Structured Memory with Temporal Reasoning for Long-Term Conversational AI ID and Graph View Contrastive Learning with Multi-View Attention Fusion for Sequential Recommendation Large Language Models to Enhance Business Process Modeling: Past, Present, and Future Trends Dual-Enhancement Product Bundling: Bridging Interactive Graph and Large Language Model Evaluation of Agents under Simulated AI Marketplace Dynamics
SIDInspector: A Mapping-First Diagnostic Resource for Sem...
[Submitted on 9 Jun 2026 (v1), last revised 11 Aug 2026 (this ve · 2026-06-09 · via cs.IR updates on arXiv.org

View PDF HTML (experimental)

Abstract:Semantic-ID (SID) tokenizers are increasingly reused as standalone artifacts in generative recommendation: an exported item-to-code mapping becomes the address space that a later sequence generator must use. These mappings rarely come with a common inspection interface, so coverage gaps, full-code aliasing, behaviorally weak prefixes, tail compression, and prefix fan-out are often found only after downstream training. We present SIDInspector, a mapping-first diagnostic resource for SID tokenizer artifacts. SIDInspector defines a small adapter contract over item mappings, metadata, interactions, and optional generator traces; validates the contract; and reports mapping-level probes for utilization, aliasing, neighborhood alignment, popularity allocation, and structural cost, with hooks for temporal churn and generator traces.
SIDInspector characterizes exported mappings independently of downstream leaderboard scores. The release covers four tokenizer lines: GRID-style RQ-KMeans under two capacity settings and a bounded ReSID/GAOQ export in a same-item Musical protocol, plus released LETTER and LC-Rec item-index artifacts. In that protocol, D2 aliasing falls from 0.977 to 0.779 under added GRID-style capacity and to 0.000 for ReSID/GAOQ, while D3 remains below a deterministic category-prefix control (0.055-0.154 versus 0.447). SIDInspector exposes capacity, addressability, and behavioral prefix alignment as distinct artifact properties before method-level evaluation. Additional domain checks show the same diagnostic distinction. Bounded fixed-reranker probes connect prefix alignment to candidate exposure, while final ranking quality remains a downstream model question.

Submission history

From: Jiandong Ding [view email]
[v1] Tue, 9 Jun 2026 03:36:31 UTC (120 KB)
[v2] Tue, 11 Aug 2026 03:09:41 UTC (115 KB)