cs.CL（2024-11-29）

📊 共 20 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) (15 🔗3) 支柱二：RL算法与架构 (RL & Architecture) (5 🔗1)

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

#题目一句话要点标签🔗⭐
1 On Domain-Adaptive Post-Training for Multimodal Large Language Models 提出领域自适应后训练方法，提升多模态大语言模型在特定领域的性能 large language model multimodal
2 Provable Scaling Laws for the Test-Time Compute of Large Language Models 提出两种可证明扩展法则的算法，提升大语言模型测试时计算效率 large language model
3 To Ensemble or Not: Assessing Majority Voting Strategies for Phishing Detection with Large Language Models 针对钓鱼URL检测，研究LLM集成投票策略的有效性 large language model
4 SDR-GNN: Spectral Domain Reconstruction Graph Neural Network for Incomplete Multimodal Learning in Conversational Emotion Recognition 提出SDR-GNN，用于会话情感识别中不完整多模态学习的光谱域重建图神经网络 multimodal
5 Can Large Language Models Reason about the Region Connection Calculus? 评估大型语言模型在区域连接演算上的推理能力 large language model
6 Ensemble Watermarks for Large Language Models 提出集成水印方法，提升大语言模型生成文本的可追溯性和抗攻击能力 large language model
7 NushuRescue: Revitalization of the Endangered Nushu Language with AI NushuRescue：利用AI技术复兴濒危女书语言 large language model
8 AIDetx: a compression-based method for identification of machine-learning generated text AIDetx：一种基于数据压缩的机器学习生成文本识别方法 large language model ✅
9 Initialization using Update Approximation is a Silver Bullet for Extremely Efficient Low-Rank Fine-Tuning LoRA-SB：通过更新近似初始化实现高效低秩微调，性能超越LoRA-XS large language model ✅
10 BatchLLM: Optimizing Large Batched LLM Inference with Global Prefix Sharing and Throughput-oriented Token Batching BatchLLM：通过全局前缀共享和吞吐量导向的Token Batching优化大规模批量LLM推理。 large language model
11 ChineseWebText 2.0: Large-Scale High-quality Chinese Web Text with Multi-dimensional and fine-grained information 构建多维度细粒度中文Web文本数据集，助力领域LLM安全研究。 large language model ✅
12 Sensitive Content Classification in Social Media: A Holistic Resource and Evaluation 提出用于社交媒体敏感内容分类的统一数据集，并验证微调LLM的有效性。 large language model
13 INCLUDE: Evaluating Multilingual Language Understanding with Regional Knowledge INCLUDE：利用区域知识评估多语言语言理解能力 large language model
14 In-Context Learning with Noisy Labels 提出噪声标签下的上下文学习任务，并设计方法提升LLM在现实场景下的泛化能力 large language model
15 KV Shifting Attention Enhances Language Modeling 提出KV移位注意力机制，提升语言模型的归纳能力和建模效率 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (5 篇)

#题目一句话要点标签🔗⭐
16 Training Agents with Weakly Supervised Feedback from Large Language Models 提出基于LLM弱监督反馈的Agent训练方法，无需专家轨迹或确定性反馈。 reinforcement learning large language model
17 LLM Teacher-Student Framework for Text Classification With No Manually Annotated Data: A Case Study in IPTC News Topic Classification 提出基于LLM的Teacher-Student框架，无需人工标注数据即可实现多语种新闻主题分类。 teacher-student large language model
18 Critical Tokens Matter: Token-Level Contrastive Estimation Enhances LLM's Reasoning Capability 提出基于关键Token对比估计的cDPO方法，提升LLM数学推理能力 DPO direct preference optimization large language model ✅
19 Reverse Thinking Makes LLMs Stronger Reasoners 提出RevThink框架，提升LLM在常识、数学和逻辑推理上的能力 distillation large language model
20 Noro: Noise-Robust One-shot Voice Conversion with Hidden Speaker Representation Learning 提出Noro，一种噪声鲁棒的单样本语音转换系统，提升噪声环境下的转换效果。 representation learning

⬅️ 返回 cs.CL 首页 · 🏠 返回主页