cs.AI（2026-08-25）

📊 共 35 篇论文 | 🔗 3 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (25 🔗2) 支柱二：RL算法与架构 (RL & Architecture) (9 🔗1) 支柱三：空间感知与语义 (Perception & Semantics) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (25 篇)

#题目一句话要点标签🔗⭐
1 SonarLLM: A Native Sonar--Optical Multimodal Large Language Model for Underwater Perception 提出SonarLLM以解决水下感知中的多模态融合问题 large language model multimodal
2 When Seeing Is Not Enough: Benchmarking Interactive Visual Grounding in LVLMs 提出交互式视觉定位框架以解决LVLMs的不足问题 visual grounding
3 Right Diagnoses, Decorative Reasoning:A Perturbation Audit of Medical Chain-of-Thought 提出医学链式思维的扰动审计以评估其可信度 chain-of-thought
4 Constraint-Guided Enterprise Data Mapping with Large Language Models 提出约束引导的企业数据映射方法以解决实体对齐问题 large language model
5 Preference Data Selection for Mitigating the Alignment Tax in Large Language Models 提出BALIGN以缓解大语言模型中的对齐税问题 large language model
6 OmniJudge or OmniBias? Diagnosing Multimodal Judges through Balanced, Decoupled Lenses 提出D3-Omni基准以解决多模态评估中的偏差问题 multimodal
7 Giraffe: A Mapping Architecture from Hidden Text Representations to Visual Embeddings for Efficient Graphic Design 提出Giraffe架构以解决多模态生成中的输入长度限制问题 large language model multimodal
8 RAGSentinel: Certifiable Geometric Consensus for Robust Retrieval-Augmented Generation 提出RAGSentinel以解决RAG系统的安全漏洞问题 large language model instruction following
9 NeuronGuard: Robust LLM Safety Alignment via Ablation-Aware Safety Signal Redistribution 提出NeuronGuard以增强大型语言模型的安全性对齐 large language model multimodal
10 ACE: A Self-Correcting Agentic Canvas Editor for Multi-Slide Presentation Automation 提出ACE以解决多幻灯片演示自动化中的布局问题 large language model instruction following
11 Relative Time Intervals Representation for Word-level Timestamping with Masked Training 提出相对时间间隔表示以解决语音模型时间戳问题 large language model TAMP
12 Evaluating Language Models on Cross-Language Code Functional Equivalence 提出PolyHuman数据集以评估跨语言代码功能等价性 large language model chain-of-thought
13 Compression Trinity: Exploring Sparsity, Quantization, and Low-Rank Approximations for LLM Compression 提出压缩三位一体框架以提升大语言模型的效率与性能 large language model
14 A Dual-Dimensional LLM Framework for Automated Item Incidental Content Similarity Analysis in Large-Scale Assessments 提出双维度框架以解决大规模评估中的内容冗余问题 large language model
15 Evidence Blindness in Direct Corpus Interaction: Persistent Navigation with AtlasNav 提出AtlasNav以解决直接语料交互中的证据盲区问题 large language model
16 Parason: Revealing Subtask and Trial Parallelism in LLM Reasoning 提出Parason以解决LLM推理中的并行性问题 large language model
17 A Literate Programming Environment for Human and Machine Agents 提出一种文艺编程环境以提升人机协作编程效率 large language model
18 When "Must" Becomes "Maybe": Constraint Weakening in LLM Agent Workflows 提出约束弱化方法以优化LLM代理工作流 large language model
19 ResiSpec: Enhancing Multi-Candidate Speculative Sampling via Residual Distribution Shaping 提出ResiSpec以解决多候选推测采样中的残差漂移问题 large language model ✅
20 Do Recipes Have Personas? Characterizing and Generating Creator Style in Attributed Procedural Graphs 提出ViralRecipesTrans以解决个体创作者风格识别问题 large language model
21 EMRB: A Multi-Level Benchmark for Evaluating LLM Reasoning over Raw Electromagnetic Signals 提出EMRB基准以评估LLM在原始电磁信号上的推理能力 large language model ✅
22 Incorporating Cognitive Load and Knowledge Transfer for Multi-Domain Knowledge Tracing 提出LT-MKT以解决多领域知识追踪中的认知负荷与知识迁移问题 large language model
23 Diverse by Reasoning: Harnessing the Wisdom of LLM Crowds for Future Prediction 提出行为感知框架以构建多样化的LLM预测群体 large language model
24 Hybrid Semantic Tool Discovery for Enterprise MCP Gateway: Architecture and Implementation 提出SCOUT以解决MCP工具发现与上下文饱和问题 large language model
25 Recursive Agentic Reasoning 提出统一视角的递归推理方法以提升模型性能 chain-of-thought

🔬 支柱二：RL算法与架构 (RL & Architecture) (9 篇)

#题目一句话要点标签🔗⭐
26 HMGCLIP: Heterogeneous Multi-Granularity Contrastive Learning for E-commerce Representation Learning 提出HMGCLIP以解决电商产品细粒度属性识别问题 representation learning contrastive learning large language model
27 Joint Optimization of Tool Creation and Use for Large Language Model Agents 提出SMITH框架以联合优化工具创建与使用 reinforcement learning large language model
28 Robust Code RL via Faulty-Code-Driven Test case Synthesis and Dense Reward Shaping 提出RobustTests框架以解决代码生成中的测试覆盖不足问题 reinforcement learning reward shaping large language model
29 PhysMLLMs: Spatial Priors for Unified Referring Segmentation and Grounded Reasoning of Images and Videos 提出PhysMLLMs以解决视频多模态模型中的时空不一致问题 distillation large language model multimodal ✅
30 AHEAD: Adaptive Hindsight with Environment-Augmented Distillation for Agentic RL 提出AHEAD框架以解决多回合RL中的监督不足问题 reinforcement learning distillation privileged information
31 Strictly Causal Streaming Video Anomaly Detection with a Theoretically-Grounded State-Space Core 提出严格因果流媒体视频异常检测方法以解决现有技术不足 Mamba SSM state space model
32 Reinforcement Learning-Guided Evolutionary Policy Optimization for Preference-Adjustable Heterogeneous Agile Earth Observation Satellite Scheduling 提出进化策略优化框架以解决异构敏捷地球观测卫星调度问题 reinforcement learning
33 OPDSearch+: On-Policy Distillation with RL Refinement for Search-Augmented Reasoning 提出OPDSearch+以解决小型语言模型的搜索增强推理问题 distillation
34 RePolicy: Reinforcement Learning for Safety-Policy Invocation in Agent Safeguards 提出RePolicy以解决语言模型代理的安全策略调用问题 reinforcement learning

🔬 支柱三：空间感知与语义 (Perception & Semantics) (1 篇)

#题目一句话要点标签🔗⭐
35 COCI: Conference Organisers and Content Identifier 提出COCI框架以解决学术会议文献结构化问题 semantic mapping semantic map large language model

⬅️ 返回 cs.AI 首页 · 🏠 返回主页