cs.CL（2024-06-13）

📊 共 42 篇论文 | 🔗 12 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (38 🔗11) 支柱二：RL算法与架构 (RL & Architecture) (4 🔗1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (38 篇)

#题目一句话要点标签🔗⭐
1 Multimodal Large Language Models with Fusion Low Rank Adaptation for Device Directed Speech Detection 提出FLoRA以解决多模态设备导向语音检测问题 large language model multimodal
2 mOSCAR: A Large-scale Multilingual and Multimodal Document-level Corpus 提出mOSCAR：一个大规模多语言多模态文档级语料库，提升多语言图像-文本任务的少样本学习能力。 large language model multimodal ✅
3 Chain of Preference Optimization: Improving Chain-of-Thought Reasoning in LLMs 提出链式偏好优化(CPO)以提升LLM的CoT推理能力 large language model chain-of-thought ✅
4 ME-Switch: A Memory-Efficient Expert Switching Framework for Large Language Models ME-Switch：面向大语言模型的高效专家切换框架，显著降低内存占用。 large language model foundation model
5 DiscreteSLU: A Large Language Model with Self-Supervised Discrete Speech Units for Spoken Language Understanding 提出DiscreteSLU，利用自监督离散语音单元增强LLM的口语理解能力 large language model instruction following
6 On the Effects of Heterogeneous Data Sources on Speech-to-Text Foundation Models OWSM v3.2：通过数据过滤和LLM增强，提升异构数据语音转文本模型的性能。 large language model foundation model
7 Living in the Moment: Can Large Language Models Grasp Co-Temporal Reasoning? 提出CoTempQA基准，评估大语言模型在并发时间推理中的能力 large language model chain-of-thought ✅
8 AlignMMBench: Evaluating Chinese Multimodal Alignment in Large Vision-Language Models AlignMMBench：首个面向中文视觉场景的多模态对齐评测基准 multimodal ✅
9 Large Language Models as Software Components: A Taxonomy for LLM-Integrated Applications 提出LLM集成应用分类法，为LLM赋能软件系统提供分析框架。 large language model
10 Understanding Jailbreak Success: A Study of Latent Space Dynamics in Large Language Models 提取Jailbreak向量以降低大语言模型越狱攻击的有效性 large language model
11 Research on Optimization of Natural Language Processing Model Based on Multimodal Deep Learning 提出基于多模态深度学习的自然语言处理模型优化方法，提升图像特征评估的鲁棒性。 multimodal
12 Multi-Modal Retrieval For Large Language Model Based Speech Recognition 提出多模态检索方法，提升基于大语言模型的语音识别性能 large language model
13 Speech ReaLLM -- Real-time Streaming Speech Recognition with Multimodal LLMs by Teaching the Flow of Time 提出Speech ReaLLM，实现基于多模态LLM的实时流式语音识别 multimodal
14 Investigating the translation capabilities of Large Language Models trained on parallel data only PLUME：仅用平行数据训练的大语言模型，探索其翻译能力。 large language model
15 SciKnowEval: Evaluating Multi-level Scientific Knowledge of Large Language Models SciKnowEval：构建多层次科学知识评估基准，衡量大语言模型科学能力 large language model
16 LLM Reading Tea Leaves: Automatically Evaluating Topic Models with Large Language Models 提出WALM：利用大语言模型自动评估主题模型，综合考量主题质量与文档表示。 large language model ✅
17 Robustness of Structured Data Extraction from In-plane Rotated Documents using Multi-Modal Large Language Models (LLM) 研究多模态LLM在倾斜文档中结构化数据提取的鲁棒性，并提出改进方向。 large language model
18 Delta-CoMe: Training-Free Delta-Compression with Mixed-Precision for Large Language Models Delta-CoMe：面向大语言模型的混合精度无训练Delta压缩 large language model
19 StructuralSleight: Automated Jailbreak Attacks on Large Language Models Utilizing Uncommon Text-Organization Structures StructuralSleight：利用罕见文本组织结构自动攻击大型语言模型 large language model
20 Enhancing Psychotherapy Counseling: A Data Augmentation Pipeline Leveraging Large Language Models for Counseling Conversations 提出一种基于LLM的数据增强流程，用于提升心理咨询对话质量 large language model ✅
21 Chain-of-Though (CoT) prompting strategies for medical error detection and correction 针对医疗错误检测与纠正，提出结合思维链（CoT）提示策略的ICL方法。 large language model chain-of-thought
22 MiLoRA: Harnessing Minor Singular Components for Parameter-Efficient LLM Finetuning MiLoRA：利用次要奇异分量进行参数高效的大语言模型微调 large language model instruction following
23 DefAn: Definitive Answer Dataset for LLMs Hallucination Evaluation DefAn：用于评估大型语言模型幻觉的权威答案数据集 large language model ✅
24 Decoding the Diversity: A Review of the Indic AI Research Landscape 综述性研究：全面解读印度语言AI研究现状与挑战 large language model
25 Navigating the Shadows: Unveiling Effective Disturbances for Modern AI Content Detectors 系统性评估AI文本检测器鲁棒性：揭示有效扰动方法与对抗学习策略 large language model ✅
26 RelevAI-Reviewer: A Benchmark on AI Reviewers for Survey Paper Relevance 提出RelevAI-Reviewer，构建AI评审基准，解决综述论文相关性评估问题 large language model
27 ReadCtrl: Personalizing text generation with readability-controlled instruction learning ReadCtrl：通过可读性控制的指令学习个性化文本生成 large language model
28 Language Models are Crossword Solvers 利用大型语言模型解决纵横填字游戏难题，显著超越现有技术水平。 large language model
29 Multi-Agent Collaboration via Cross-Team Orchestration 提出Croto，通过跨团队协作编排提升LLM驱动的智能体在复杂任务中的表现。 large language model ✅
30 CLST: Cold-Start Mitigation in Knowledge Tracing by Aligning a Generative Language Model as a Students' Knowledge Tracer CLST：通过对齐生成式语言模型，缓解知识追踪中的冷启动问题 large language model
31 An Approach to Build Zero-Shot Slot-Filling System for Industry-Grade Conversational Assistants 提出一种基于小型LLM的零样本槽填充系统，用于工业级对话助手。 large language model
32 Newswire: A Large-Scale Structured Database of a Century of Historical News 构建大规模历史新闻数据库Newswire，助力语言模型和社会科学研究。 large language model
33 Sharing Matters: Analysing Neurons Across Languages and Tasks in LLMs 分析LLM跨语言和任务的神经元共享模式，揭示多语言模型内部机制 large language model
34 Test of Time: A Benchmark for Evaluating LLMs on Temporal Reasoning 提出Test of Time基准，评估LLM在时间推理上的能力 large language model ✅
35 Bayesian Statistical Modeling with Predictors from LLMs 利用LLM预测器进行贝叶斯统计建模，评估其人类行为预测能力 large language model
36 Plan, Generate and Complicate: Improving Low-resource Dialogue State Tracking via Easy-to-Difficult Zero-shot Data Augmentation 提出EDZ-DA框架，通过由易到难的零样本数据增强提升低资源对话状态跟踪性能。 large language model
37 StreamBench: Towards Benchmarking Continuous Improvement of Language Agents StreamBench：面向语言智能体持续改进的评测基准 large language model ✅
38 Standard Language Ideology in AI-Generated Language 揭示大型语言模型中标准语言意识形态，强调其对少数语言社区的影响。 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🔗⭐
39 Unpacking DPO and PPO: Disentangling Best Practices for Learning from Preference Feedback 系统性解耦偏好学习各因素，揭示数据质量、算法选择等对语言模型性能的影响 PPO DPO instruction following ✅
40 Mixture-of-Skills: Learning to Optimize Data Usage for Fine-Tuning Large Language Models 提出混合技能（MoS）框架，通过强化学习优化LLM微调中的数据使用。 reinforcement learning large language model
41 ContraSolver: Self-Alignment of Language Models by Resolving Internal Preference Contradictions ContraSolver：通过解决内部偏好矛盾实现语言模型的自对齐 DPO direct preference optimization large language model
42 Modeling Comparative Logical Relation with Contrastive Learning for Text Generation 提出CoLo模型，利用对比学习建模比较逻辑关系，用于数据到文本生成任务。 contrastive learning

⬅️ 返回 cs.CL 首页 · 🏠 返回主页