cs.AI（2024-05-31）

📊 共 16 篇论文 | 🔗 2 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (12 🔗2) 支柱二：RL算法与架构 (RL & Architecture) (3) 支柱五：交互与反应 (Interaction & Reaction) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (12 篇)

#题目一句话要点标签🔗⭐
1 ToxVidLM: A Multimodal Framework for Toxicity Detection in Code-Mixed Videos ToxVidLM：提出一种多模态框架，用于检测混合语视频中的有害内容。 multimodal
2 Large Language Models are Zero-Shot Next Location Predictors 利用大型语言模型实现零样本的下一位置预测 large language model
3 Unraveling and Mitigating Retriever Inconsistencies in Retrieval-Augmented Large Language Models 提出检索器集成框架EoR，解决检索增强大语言模型中检索不一致性问题 large language model
4 Robust Planning with LLM-Modulo Framework: Case Study in Travel Planning LLM-Modulo框架提升LLM在旅行规划任务中的鲁棒性 large language model chain-of-thought
5 Grammar-Aligned Decoding 提出语法对齐解码ASAp，解决LLM在语法约束下生成低质量结构化输出的问题。 large language model
6 RAG Does Not Work for Enterprises 针对企业级应用，分析RAG的局限性并提出评估框架 large language model
7 LLM-RankFusion: Mitigating Intrinsic Inconsistency in LLM-based Ranking 提出LLM-RankFusion框架，缓解LLM排序中固有的不一致性问题。 large language model ✅
8 Standards for Belief Representations in LLMs 为大型语言模型中的信念表征研究，提出一套评估标准体系。 large language model
9 Enhancing Noise Robustness of Retrieval-Augmented Language Models with Adaptive Adversarial Training 提出RAAT：一种基于自适应对抗训练的检索增强语言模型噪声鲁棒性提升方法 large language model ✅
10 A Robot Walks into a Bar: Can Language Models Serve as Creativity Support Tools for Comedy? An Evaluation of LLMs' Humour Alignment with Comedians 评估大型语言模型在喜剧创作中的创造力支持能力，揭示其偏见与局限性 large language model
11 No Free Lunch Theorem for Privacy-Preserving LLM Inference 针对隐私保护LLM推理，提出NFL定理以分析隐私与效用间的权衡 large language model
12 Generative AI Voting: Fair Collective Choice is Resilient to LLM Biases and Inconsistencies 利用生成式AI投票提升直接民主韧性：公平集体选择抵御LLM偏差与不一致性 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (3 篇)

#题目一句话要点标签🔗⭐
13 Mind the Inconspicuous: Revealing the Hidden Weakness in Aligned LLMs' Refusal Boundaries 揭示对齐LLM拒绝边界的隐蔽弱点：EOS token引发上下文分割 reinforcement learning RLHF large language model
14 OpenTensor: Reproducing Faster Matrix Multiplication Discovering Algorithms OpenTensor：复现并加速矩阵乘法算法发现，提升计算效率 reinforcement learning deep reinforcement learning DRL
15 ADESSE: Advice Explanations in Complex Repeated Decision-Making Environments ADESSE：在复杂重复决策环境中提供可解释建议，提升人机协作。 reinforcement learning deep reinforcement learning

🔬 支柱五：交互与反应 (Interaction & Reaction) (1 篇)

#题目一句话要点标签🔗⭐
16 Locking Machine Learning Models into Hardware 提出一种硬件锁定的机器学习模型保护方案，防止未经授权的使用。 OMOMO

⬅️ 返回 cs.AI 首页 · 🏠 返回主页