cs.AI（2024-09-26）

📊 共 31 篇论文 | 🔗 4 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (21 🔗3) 支柱二：RL算法与架构 (RL & Architecture) (8) 支柱七：动作重定向 (Motion Retargeting) (1 🔗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) (21 篇)

#题目一句话要点标签🔗⭐
1 MMMT-IF: A Challenging Multimodal Multi-Turn Instruction Following Benchmark 提出MMMT-IF基准测试，用于评估多模态多轮对话中指令遵循能力。 multimodal instruction following
2 Compositional Hardness of Code in Large Language Models -- A Probabilistic Perspective 揭示LLM代码生成中组合任务的内在困难，提出多智能体分解策略 large language model chain-of-thought
3 Learning to Love Edge Cases in Formative Math Assessment: Using the AMMORE Dataset and Chain-of-Thought Prompting to Improve Grading Accuracy 提出AMMORE数据集，并利用CoT提示提升LLM在数学形成性评估中边缘案例的评分准确率 large language model chain-of-thought
4 Harmful Fine-tuning Attacks and Defenses for Large Language Models: A Survey 综述有害微调攻击与防御，应对大语言模型安全风险 large language model ✅
5 A Multimodal Single-Branch Embedding Network for Recommendation in Cold-Start and Missing Modality Scenarios 提出SiBraR单分支嵌入网络，解决推荐系统中冷启动和模态缺失问题。 multimodal
6 MaskLLM: Learnable Semi-Structured Sparsity for Large Language Models MaskLLM：面向大语言模型的可学习半结构化稀疏方法 large language model ✅
7 Development and Validation of a Large Language Model for Generating Fully-Structured Radiology Reports 提出动态模板约束解码的LLM，用于生成高质量、结构化的肺癌筛查报告。 large language model
8 Trustworthy AI: Securing Sensitive Data in Large Language Models 提出面向大语言模型的信任框架，保障敏感数据安全 large language model
9 A Scalable Data-Driven Framework for Systematic Analysis of SEC 10-K Filings Using Large Language Models 提出一种可扩展的数据驱动框架，利用大型语言模型系统分析SEC 10-K文件。 large language model
10 Infer Human's Intentions Before Following Natural Language Instructions 提出FISER框架，通过社交推理预测人类意图，提升具身协作任务中的指令跟随性能。 instruction following chain-of-thought
11 A Survey of Spatio-Temporal EEG data Analysis: from Models to Applications 综述时空脑电图数据分析方法及其应用 large language model foundation model ✅
12 Policy Maps: Tools for Guiding the Unbounded Space of LLM Behaviors 提出Policy Maps，引导LLM行为空间，辅助AI策略设计。 large language model
13 Data-Prep-Kit: getting your data ready for LLM application development 提出Data Prep Kit (DPK)，用于大规模语言模型应用开发的数据准备。 large language model
14 MoJE: Mixture of Jailbreak Experts, Naive Tabular Classifiers as Guard for Prompt Attacks 提出MoJE：一种基于专家混合和朴素表格分类器的LLM越狱攻击防御方法 large language model
15 Heuristics and Biases in AI Decision-Making: Implications for Responsible AGI 评估LLM认知偏差：揭示GPT-4o、Gemma 2和Llama 3.1的决策缺陷 large language model
16 The Nexus of AR/VR, AI, UI/UX, and Robotics Technologies in Enhancing Learning and Social Interaction for Children with Autism Spectrum Disorders: A Systematic Review 系统综述：AR/VR、AI、UI/UX与机器人技术融合，提升自闭症儿童的学习与社交互动 large language model
17 AI Delegates with a Dual Focus: Ensuring Privacy and Strategic Self-Disclosure 提出双重关注的AI代理，保障隐私与策略性自我披露，用于社交互动。 large language model
18 Dr. GPT in Campus Counseling: Understanding Higher Education Students' Opinions on LLM-assisted Mental Health Services 探索LLM在校园心理咨询中的应用：理解大学生对AI辅助心理健康服务的观点 large language model
19 Multi-Designated Detector Watermarking for Language Models 提出多指定检测器水印（MDDW）技术，用于保护大型语言模型的知识产权。 large language model
20 From News to Forecast: Integrating Event Analysis in LLM-Based Time Series Forecasting with Reflection 提出融合事件分析与LLM的时间序列预测方法，提升预测精度。 large language model
21 Human Mobility Modeling with Household Coordination Activities under Limited Information via Retrieval-Augmented LLMs 提出检索增强LLM框架，利用有限信息建模包含家庭协同的人类出行模式 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (8 篇)

#题目一句话要点标签🔗⭐
22 Navigation in a simplified Urban Flow through Deep Reinforcement Learning 提出基于PPO+LSTM的深度强化学习方法，优化无人机在城市环境中的自主导航。 reinforcement learning deep reinforcement learning DRL
23 Hierarchical End-to-End Autonomous Driving: Integrating BEV Perception with Deep Reinforcement Learning 提出结合BEV感知与深度强化学习的端到端自动驾驶框架，提升驾驶性能。 reinforcement learning deep reinforcement learning DRL
24 DRL-STNet: Unsupervised Domain Adaptation for Cross-modality Medical Image Segmentation via Disentangled Representation Learning DRL-STNet：通过解耦表征学习实现跨模态医学图像分割的无监督域自适应 DRL representation learning
25 Role-RL: Online Long-Context Processing with Role Reinforcement Learning for Distinct LLMs in Their Optimal Roles 提出Role-RL，通过角色强化学习实现LLM在线长文本处理中的最优角色分配 reinforcement learning large language model
26 FactorSim: Generative Simulation via Factorized Representation FactorSim：通过分解表示生成模拟环境，用于训练智能体。 reinforcement learning zero-shot transfer
27 Autonomous Network Defence using Reinforcement Learning 提出基于强化学习的自主网络防御方法，有效应对高级持续性威胁 reinforcement learning
28 Breaking PEFT Limitations: Leveraging Weak-to-Strong Knowledge Transfer for Backdoor Attacks in LLMs 提出FAKD：利用弱到强知识迁移增强LLM中基于PEFT的后门攻击 distillation large language model
29 Just Say What You Want: Only-prompting Self-rewarding Online Preference Optimization 提出仅提示自奖励在线偏好优化算法，提升小模型在线RLHF性能。 reinforcement learning RLHF

🔬 支柱七：动作重定向 (Motion Retargeting) (1 篇)

#题目一句话要点标签🔗⭐
30 A Time Series is Worth Five Experts: Heterogeneous Mixture of Experts for Traffic Flow Prediction 提出异构混合专家模型TITAN，用于解决交通流量预测中变量中心学习不足的问题。 spatial relationship ✅

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

#题目一句话要点标签🔗⭐
31 Subjective and Objective Quality-of-Experience Evaluation Study for Live Video Streaming 针对直播视频流提出主客观QoE评估方法，并构建TaoLive QoE数据集。 optical flow

⬅️ 返回 cs.AI 首页 · 🏠 返回主页