cs.AI（2025-04-19）

📊 共 19 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) (15 🔗2) 支柱二：RL算法与架构 (RL & Architecture) (2 🔗1) 支柱八：物理动画 (Physics-based Animation) (1) 支柱一：机器人控制 (Robot Control) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (15 篇)

#	题目	一句话要点	标签	🔗	⭐
1	Improving the Serving Performance of Multi-LoRA Large Language Models via Efficient LoRA and KV Cache Management	FASTLIBRA：通过高效LoRA和KV缓存管理提升多LoRA大语言模型服务性能	large language model
2	Large Language Model Enhanced Particle Swarm Optimization for Hyperparameter Tuning for Deep Learning Models	提出LLM增强的粒子群优化算法，用于深度学习模型超参数调优，显著提升收敛速度和计算效率。	large language model
3	Expanding the Generative AI Design Space through Structured Prompting and Multimodal Interfaces	ACAI：通过结构化提示和多模态界面扩展生成式AI在广告设计中的应用	multimodal
4	EmbedAgent: Benchmarking Large Language Models in Embedded System Development	EmbedAgent：评估大语言模型在嵌入式系统开发中的能力	large language model
5	A Data-Centric Approach for Safe and Secure Large Language Models against Threatening and Toxic Content	提出基于BART的后生成修正模型，提升大语言模型应对威胁和有害内容的安全性和可靠性。	large language model
6	PipeWeaver: Addressing Data Dynamicity in Large Multimodal Model Training with Dynamic Interleaved Pipeline	PipeWeaver：通过动态交错流水线解决大规模多模态模型训练中的数据动态性问题	multimodal
7	CodeCrash: Exposing LLM Fragility to Misleading Natural Language in Code Reasoning	CodeCrash：揭示LLM在代码推理中对误导性自然语言的脆弱性	large language model chain-of-thought
8	The Geometry of Self-Verification in a Task-Specific Reasoning Model	研究任务型推理模型中的自验证几何结构，揭示模型如何验证自身答案。	chain-of-thought
9	SOTOPIA-S4: a user-friendly system for flexible, customizable, and large-scale social simulation	SOTOPIA-S4：一个灵活、可定制的大规模社会模拟用户友好系统	large language model
10	An Empirical Study of LLM Reasoning Ability Under Strict Output Length Constraint	研究LLM在严格输出长度约束下的推理能力，揭示模型选择和提示策略的权衡。	large language model
11	LLM-Enhanced Black-Litterman Portfolio Optimization	提出LLM增强的Black-Litterman模型，利用LLM预测指导投资组合优化。	large language model	✅
12	FAIRGAME: a Framework for AI Agents Bias Recognition using Game Theory	FAIRGAME：一个利用博弈论识别AI智能体偏差的框架	large language model
13	RINN: One Sample Radio Frequency Imaging based on Physics Informed Neural Network	提出RINN：一种基于物理信息神经网络的单样本射频成像方法	multimodal
14	Explainable Recommendation with Simulated Human Feedback	提出基于模拟人类反馈的可解释推荐框架，解决现有方法缺乏有效反馈信号的问题。	large language model
15	TALES: Text Adventure Learning Environment Suite	TALES：用于评估LLM推理能力的文本冒险学习环境套件	large language model	✅

🔬 支柱二：RL算法与架构 (RL & Architecture) (2 篇)

#	题目	一句话要点	标签	🔗	⭐
16	InfiGUI-R1: Advancing Multimodal GUI Agents from Reactive Actors to Deliberative Reasoners	InfiGUI-R1：提出Actor2Reasoner框架，提升多模态GUI Agent的推理能力。	reinforcement learning distillation large language model	✅
17	Direct Advantage Regression: Aligning LLMs with Online AI Reward	提出直接优势回归(DAR)，利用在线AI奖励对齐LLM，提升人机一致性。	reinforcement learning RLHF

🔬 支柱八：物理动画 (Physics-based Animation) (1 篇)

#	题目	一句话要点	标签	🔗	⭐
18	Data Augmentation Using Neural Acoustic Fields With Retrieval-Augmented Pre-training	提出基于检索增强预训练的神经声场数据增强方法，用于房间脉冲响应估计。	PULSE

🔬 支柱一：机器人控制 (Robot Control) (1 篇)

#	题目	一句话要点	标签	🔗	⭐
19	ScholarMate: A Mixed-Initiative Tool for Qualitative Knowledge Work and Information Sensemaking	ScholarMate：一种混合主动的工具，用于定性知识工作和信息理解。	manipulation

⬅️ 返回 cs.AI 首页 · 🏠 返回主页