cs.AI（2025-02-06）

📊 共 16 篇论文

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (10) 支柱二：RL算法与架构 (RL & Architecture) (4) 支柱八：物理动画 (Physics-based Animation) (2)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (10 篇)

#题目一句话要点标签🔗⭐
1 A Survey on Backdoor Threats in Large Language Models (LLMs): Attacks, Defenses, and Evaluations 综述大型语言模型（LLMs）中的后门威胁：攻击、防御与评估 large language model
2 Assessing and Prioritizing Ransomware Risk Based on Historical Victim Data 提出基于历史受害者数据的勒索软件风险评估与优先级排序方法 large language model chain-of-thought
3 UniForm: A Unified Multi-Task Diffusion Transformer for Audio-Video Generation UniForm：用于音视频生成的统一多任务扩散Transformer large language model
4 Fine, I'll Merge It Myself: A Multi-Fidelity Framework for Automated Model Merging 提出多保真度模型自动融合框架，提升大语言模型推理能力。 large language model
5 Safety is Essential for Responsible Open-Ended Systems 强调开放式AI系统安全性的重要性，呼吁关注并解决潜在风险。 foundation model
6 ADIFF: Explaining audio difference using natural language ADIFF：提出一种利用自然语言解释音频差异的框架与基准。 large language model
7 Llasa: Scaling Train-Time and Inference-Time Compute for Llama-based Speech Synthesis Llasa：通过扩展训练和推理计算资源，提升基于Llama的语音合成效果 large language model
8 VTutor: An Open-Source SDK for Generative AI-Powered Animated Pedagogical Agents with Multi-Media Output VTutor：一个开源SDK，用于构建基于生成式AI的多媒体动画教学代理 large language model
9 Automating a Complete Software Test Process Using LLMs: An Automotive Case Study 提出基于LLM的自动化软件测试系统，解决车载API测试难题 large language model
10 Understanding and Supporting Formal Email Exchange by Answering AI-Generated Questions 提出基于问答的LLM辅助邮件回复系统ResQ，提升回复效率并降低认知负荷。 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (4 篇)

#题目一句话要点标签🔗⭐
11 QExplorer: Large Language Model Based Query Extraction for Toxic Content Exploration QExplorer：基于大语言模型的查询提取方法，用于探索有害内容 DPO direct preference optimization large language model
12 Boosting Knowledge Graph-based Recommendations through Confidence-Aware Augmentation with Large Language Models 提出CKG-LLMA框架，通过置信度感知增强知识图谱，提升推荐系统性能。 contrastive learning large language model
13 Dark Distillation: Backdooring Distilled Datasets without Accessing Raw Data 提出Dark Distillation，实现无需原始数据即可对蒸馏数据集进行后门攻击。 distillation
14 Free Energy Risk Metrics for Systemically Safe AI: Gatekeeping Multi-Agent Study 提出基于自由能原理的累积风险暴露度量，用于提升多智能体系统安全性。 world model penetration

🔬 支柱八：物理动画 (Physics-based Animation) (2 篇)

#题目一句话要点标签🔗⭐
15 A Hybrid Model for Weakly-Supervised Speech Dereverberation 提出一种混合模型，利用弱监督提升语音去混响系统性能 PULSE
16 Multi-Objective Mobile Damped Wave Algorithm (MOMDWA): A Novel Approach For Quantum System Control 提出多目标移动阻尼波算法(MOMDWA)，用于提升量子系统控制的效率与鲁棒性。 PULSE

⬅️ 返回 cs.AI 首页 · 🏠 返回主页