Flagship large language models (LLMs) offer rich general knowledge and strong reasoning capabilities, but their high inference cost and latency limit their use for large-scale data labeling. We study zero-shot text classification with semantically meaningful labels and task descriptions, without task-specific labeled examples or classifier feedback. We propose \textbf{SLKR} (Semantic Label Knowledge Retrieval), a label-centric retrieval-augmented generation framework that separates offline knowledge construction from lightweight online inference. Offline, a flagship LLM independently generates label definitions, retrieval keywords, and boundary rules for potentially confusable label pairs, forming a reusable label-knowledge store. Because this construction operates over the label space rather than the input corpus, its one-time flagship cost can be amortized across large-scale inference. Online, hybrid semantic and lexical retrieval first identifies the top-$K$ candidate labels and then selects up to $M$ input-relevant boundary rules connecting candidate-label pairs. A lightweight LLM performs the final classification in a single call using the retrieved definitions and rules. Across three public benchmarks and a real-world multilingual book-classification dataset, SLKR achieves the best results on all four datasets with Qwen3.5-4B and on three of four with Ministral-3-3B, while using compact input contexts and no task-specific labeled examples. Boundary ablations further show that retrieving a small set of input-relevant rules preserves performance close to using all boundaries with substantially less context, supporting efficient reuse of flagship-generated label knowledge for lightweight classification.
R3G: A Reasoning-Retrieval-Reranking Framework for Vision-Centric Answer Generation
The Shenzhen International Graduate School, Tsinghua University, China(清华大学深圳国际研究生院);State Key Laboratory of Nuclear Power Safety Technology and Equipment, China(核能安全技术与装备国家重点实验室);School of Computer Science and Information Engineering, Hefei University of Technology, China(合肥工业大学计算机科学与信息工程学院)
StoryTR: Narrative-Centric Video Temporal Retrieval with Theory of Mind Reasoning
Department of Computer Science and Engineering, The Chinese University of Hong Kong, Hong Kong(香港中文大学计算机科学与工程系);SkyReels, Skywork AI(SkyReels与Skywork AI);Independent Researcher(独立研究者)
Query-centric Audio-Visual Cognition Network for Moment Retrieval, Segmentation and Step-Captioning
University of Chinese Academy of Sciences(中国科学院大学);Institute of Computing Technology, Chinese Academy of Sciences(中国科学院计算技术研究所);Peng Cheng Laboratory(鹏城实验室)
IRSC: A Zero-shot Evaluation Benchmark for Information Retrieval through Semantic Comprehension in Retrieval-Augmented Generation Scenarios
PengCheng Laboratory(鹏城实验室);Shenzhen International Graduate School, Tsinghua University(清华大学深圳国际研究生院);Southern University of Science and Technology(南方科技大学)
A Rationale-Centric Framework for Human-in-the-loop Machine Learning
The Insight Centre for Data Analytics, University College Dublin(都柏林大学洞察数据分析中心);SenseTime Research(商汤科技研究院);Westlake University(西湖大学);Westlake Institute for Advanced Study(西湖高等研究院)
Geometric-Guided Few-Shot Dental Landmark Detection with Human-Centric Foundation Model
Department of Electronic and Computer Engineering, The Hong Kong University of Science and Technology, Hong Kong SAR, China(电子与计算机工程系,香港科学与技术大学);Division of Paediatric Dentistry and Orthodontics, Faculty of Dentistry, The University of Hong Kong, Hong Kong SAR, China(牙科学院儿童牙科与正畸学系,香港大学)