LLM-Tool-Survey

LLM-Tool-Survey

大型语言模型工具学习调查研究

该研究系统性调查大型语言模型(LLMs)通过工具学习增强解决复杂问题能力。从工具学习的优势和实现方法两方面全面回顾现有文献,总结基准测试和评估方法,讨论当前挑战和未来方向,为相关研究和开发提供见解。

大语言模型工具学习人工智能综述自然语言处理Github开源项目

Survey: Tool Learning with Large Language Models

Recently, tool learning with large language models(LLMs) has emerged as a promising paradigm for augmenting the capabilities of LLMs to tackle highly complex problems.

This is the collection of papers related to tool learning with LLMs. These papers are organized according to our survey paper "Tool Learning with Large Language Models: A Survey".

中文: We have noticed that PaperAgent and 旺知识 have provided a brief and a comprehensive introduction in Chinese, respectively. We greatly appreciate their assistance.

Please feel free to contact us if you have any questions or suggestions!

Contribution

:tada::+1: Please feel free to open an issue or make a pull request! :tada::+1:

Citation

If you find our work helps your research, please kindly cite our paper:

@article{qu2024toolsurvey, author={Qu, Changle and Dai, Sunhao and Wei, Xiaochi and Cai, Hengyi and Wang, Shuaiqiang and Yin, Dawei and Xu, Jun and Wen, Ji-Rong}, title={Tool Learning with Large Language Models: A Survey}, journal={arXiv preprint arXiv:2405.17935}, year={2024} }

📋 Contents

🌟 Introduction

Recently, tool learning with large language models (LLMs) has emerged as a promising paradigm for augmenting the capabilities of LLMs to tackle highly complex problems. Despite growing attention and rapid advancements in this field, the existing literature remains fragmented and lacks systematic organization, posing barriers to entry for newcomers. This gap motivates us to conduct a comprehensive survey of existing works on tool learning with LLMs. In this survey, we focus on reviewing existing literature from the two primary aspects (1) why tool learning is beneficial and (2) how tool learning is implemented, enabling a comprehensive understanding of tool learning with LLMs. We first explore the “why” by reviewing both the benefits of tool integration and the inherent benefits of the tool learning paradigm from six specific aspects. In terms of “how”, we systematically review the literature according to a taxonomy of four key stages in the tool learning workflow: task planning, tool selection, tool calling, and response generation. Additionally, we provide a detailed summary of existing benchmarks and evaluation methods, categorizing them according to their relevance to different stages. Finally, we discuss current challenges and outline potential future directions, aiming to inspire both researchers and industrial developers to further explore this emerging and promising area.

The overall workflow for tool learning with large language models.

<div align=center> <img src="assets/Framework.png" height="500"/> </div>

📄 Paper List

Why Tool Learning?

Benefit of Tools.

  • Knowledge Acquisition.

    • Search Engine

      Internet-Augmented Dialogue Generation, ACL 2022. [Paper]

      WebGPT: Browser-assisted question-answering with human feedback, Preprint 2021. [Paper]

      Internet-augmented language models through few-shot prompting for open-domain question answering, Preprint 2022. [Paper]

      REPLUG: Retrieval-Augmented Black-Box Language Models, Preprint 2023. [Paper]

      Toolformer: Language Models Can Teach Themselves to Use Tools, NeurIPS 2023. [Paper]

      ART: Automatic multi-step reasoning and tool-use for large language models, Preprint 2023. [Paper]

      ToolCoder: Teach Code Generation Models to use API search tools, Preprint 2023. [Paper]

      CRITIC: Large Language Models Can Self-Correct with Tool-Interactive Critiquing, ICLR 2024. [Paper]

    • Database & Knowledge Graph

      Lamda: Language models for dialog applications, Preprint 2022. [Paper]

      Gorilla: Large Language Model Connected with Massive APIs, Preprint 2023. [Paper]

      ToolkenGPT: Augmenting Frozen Language Models with Massive Tools via Tool Embeddings, NeurIPS 2023. [Paper]

      ToolQA: A Dataset for LLM Question Answering with External Tools, NeurIPS 2023. [Paper]

      Syntax Error-Free and Generalizable Tool Use for LLMs via Finite-State Decoding, NeurIPS 2023. [Paper]

      Middleware for LLMs: Tools are Instrumental for Language Agents in Complex Environments, Preprint 2024. [Paper]

    • Weather or Map

      On the Tool Manipulation Capability of Open-source Large Language Models, NeurIPS 2023. [Paper]

      ToolAlpaca: Generalized Tool Learning for Language Models with 3000 Simulated Cases, Preprint 2023. [Paper]

      Tool Learning with Foundation Models, Preprint 2023. [Paper]

  • Expertise Enhancement.

    • Mathematical Tools

      Training verifiers to solve math word problems, Preprint 2021. [Paper]

      MRKL Systems: A modular, neuro-symbolic architecture that combines large language models, external knowledge sources and discrete reasoning, Preprint 2021. [Paper]

      Chaining Simultaneous Thoughts for Numerical Reasoning, EMNLP 2022. [Paper]

      Calc-X and Calcformers: Empowering Arithmetical Chain-of-Thought through Interaction with Symbolic Systems, EMNLP 2023. [Paper]

      Solving math word problems by combining language models with symbolic solvers, NeurIPS 2023. [Paper]

      Evaluating and improving tool-augmented computation-intensive math reasoning, NeurIPS 2023. [Paper]

      ToRA: A Tool-Integrated Reasoning Agent for Mathematical Problem Solving, ICLR 2024. [Paper]

      MATHSENSEI: A Tool-Augmented Large Language Model for Mathematical Reasoning, Preprint 2024. [Paper]

      Calc-CMU at SemEval-2024 Task 7: Pre-Calc -- Learning to Use the Calculator Improves Numeracy in Language Models, NAACL 2024. [Paper]

      MathViz-E: A Case-study in Domain-Specialized Tool-Using Agents, Preprint 2024. [Paper]

    • Python Interpreter

      Pal: Program-aided language models, ICML 2023. [Paper]

      Program of Thoughts Prompting: Disentangling Computation from Reasoning for Numerical Reasoning Tasks, TMLR 2023. [Paper]

      Fact-Checking Complex Claims with Program-Guided Reasoning, ACL 2023. [Paper]

      Chameleon: Plug-and-Play Compositional Reasoning with Large Language Models, NeurIPS 2023. [Paper]

      LeTI: Learning to Generate from Textual Interactions, NAACL 2024. [Paper]

      Mint: Evaluating llms in multi-turn interaction with tools and language feedback, ICLR 2024. [Paper]

      Executable Code Actions Elicit Better LLM Agents, ICML 2024. [Paper]

      CodeNav: Beyond tool-use to using real-world codebases with LLM agents, Preprint 2024. [Paper]

      APPL: A Prompt Programming Language for Harmonious Integration of Programs and Large Language Model Prompts, Preprint 2024. [Paper]

      BigCodeBench: Benchmarking Code Generation with Diverse Function Calls and Complex Instructions, Preprint 2024. [Paper]

      CodeAgent: Enhancing Code Generation with Tool-Integrated Agent Systems for Real-World Repo-level Coding Challenges, ACL 2024. [Paper]

    • Others

      Chemical: MultiTool-CoT: GPT-3 Can Use Multiple External Tools with Chain of Thought Prompting, ACL 2023. [Paper]

      ChemCrow: Augmenting large-language models with chemistry tools, Nature Machine Intelligence 2024. [Paper]

      A REVIEW OF LARGE LANGUAGE MODELS AND AUTONOMOUS AGENTS IN CHEMISTRY, Preprint 2024. [Paper]

      Biomedical: GeneGPT: Augmenting Large Language Models with Domain Tools for Improved Access to Biomedical Information, ISMB 2024. [Paper]

      Financial: Equipping Language Models with Tool Use Capability for Tabular Data Analysis in Finance, EACL 2024. [Paper]

      Financial: Simulating Financial Market via Large Language Model based Agents, Preprint 2024. [Paper]

      Medical: AgentMD: Empowering Language Agents for Risk Prediction with Large-Scale Clinical Tool Learning, Preprint 2024. [Paper]

      MMedAgent: Learning to Use Medical Tools with Multi-modal Agent, Preprint 2024. [Paper]

      Recommendation: Let Me Do It For You: Towards LLM Empowered Recommendation via Tool Learning, SIGIR 2024. [Paper]

      Gas Turbines: DOMAIN-SPECIFIC ReAct FOR PHYSICS-INTEGRATED ITERATIVE MODELING: A CASE STUDY OF LLM AGENTS FOR GAS PATH ANALYSIS OF GAS TURBINES, Preprint 2024. [Paper]

      WORLDAPIS: The World Is Worth How Many APIs? A Thought Experiment, ACL 2024 Workshop. [Paper]

  • Automation and Efficiency.

    • Schedule Tools

      ToolQA: A Dataset for LLM Question Answering with External Tools, NeurIPS 2023. [Paper]

    • Set Reminders

      ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIs, ICLR 2024. [Paper]

    • Filter Emails

      ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIs, ICLR 2024. [Paper]

    • Project Management

      ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIs, ICLR 2024. [Paper]

    • Online Shopping Assistants

      WebShop: Towards Scalable Real-World Web Interaction with Grounded Language Agents, NeurIPS 2022. [Paper]

  • Interaction Enhancement.

    • Multi-modal Tools

      Vipergpt: Visual inference via python execution for reasoning, ICCV 2023. [Paper]

      MM-REACT: Prompting ChatGPT for Multimodal Reasoning and Action, Preprint 2023. [Paper]

      InternGPT: Solving Vision-Centric Tasks by Interacting with ChatGPT Beyond Language, Preprint 2023. [Paper]

      AssistGPT: A General Multi-modal Assistant that can Plan, Execute, Inspect, and Learn, Preprint 2023. [Paper]

      CLOVA: A closed-loop visual assistant with tool usage and update, CVPR 2024. [Paper]

      DiffAgent: Fast and Accurate Text-to-Image API Selection with Large Language Model, CVPR 2024. [Paper]

      MLLM-Tool: A Multimodal Large Language Model For Tool Agent Learning, Preprint 2024. [Paper]

      m&m's: A Benchmark to Evaluate Tool-Use for multi-step multi-modal Tasks, Preprint 2024. [Paper]

      From the Least to the Most: Building a Plug-and-Play Visual Reasoner via Data Synthesis, Preprint 2024. [Paper]

    • Machine Translator

      Toolformer: Language Models Can Teach Themselves to Use Tools, NeurIPS 2023. [Paper]

      Tool Learning with Foundation Models, Preprint 2023. [Paper]

    • Natural Language Processing Tools

      HuggingGPT: Solving AI Tasks with ChatGPT and its Friends in Hugging Face, NeurIPS 2023. [Paper]

      GitAgent: Facilitating Autonomous Agent with GitHub by Tool Extension, Preprint 2023. [Paper]

Benefit of Tool Learning.

  • Enhanced Interpretability and User Trust.
  • Improved Robustness and Adaptability.

How

编辑推荐精选

AEE

AEE

AI Excel全自动制表工具

AEE 在线 AI 全自动 Excel 编辑器,提供智能录入、自动公式、数据整理、图表生成等功能,高效处理 Excel 任务,提升办公效率。支持自动高亮数据、批量计算、不规则数据录入,适用于企业、教育、金融等多场景。

UI-TARS-desktop

UI-TARS-desktop

基于 UI-TARS 视觉语言模型的桌面应用,可通过自然语言控制计算机进行多模态操作。

UI-TARS-desktop 是一款功能强大的桌面应用,基于 UI-TARS(视觉语言模型)构建。它具备自然语言控制、截图与视觉识别、精确的鼠标键盘控制等功能,支持跨平台使用(Windows/MacOS),能提供实时反馈和状态显示,且数据完全本地处理,保障隐私安全。该应用集成了多种大语言模型和搜索方式,还可进行文件系统操作。适用于需要智能交互和自动化任务的场景,如信息检索、文件管理等。其提供了详细的文档,包括快速启动、部署、贡献指南和 SDK 使用说明等,方便开发者使用和扩展。

Wan2.1

Wan2.1

开源且先进的大规模视频生成模型项目

Wan2.1 是一个开源且先进的大规模视频生成模型项目,支持文本到图像、文本到视频、图像到视频等多种生成任务。它具备丰富的配置选项,可调整分辨率、扩散步数等参数,还能对提示词进行增强。使用了多种先进技术和工具,在视频和图像生成领域具有广泛应用前景,适合研究人员和开发者使用。

爱图表

爱图表

全流程 AI 驱动的数据可视化工具,助力用户轻松创作高颜值图表

爱图表(aitubiao.com)就是AI图表,是由镝数科技推出的一款创新型智能数据可视化平台,专注于为用户提供便捷的图表生成、数据分析和报告撰写服务。爱图表是中国首个在图表场景接入DeepSeek的产品。通过接入前沿的DeepSeek系列AI模型,爱图表结合强大的数据处理能力与智能化功能,致力于帮助职场人士高效处理和表达数据,提升工作效率和报告质量。

Qwen2.5-VL

Qwen2.5-VL

一款强大的视觉语言模型,支持图像和视频输入

Qwen2.5-VL 是一款强大的视觉语言模型,支持图像和视频输入,可用于多种场景,如商品特点总结、图像文字识别等。项目提供了 OpenAI API 服务、Web UI 示例等部署方式,还包含了视觉处理工具,有助于开发者快速集成和使用,提升工作效率。

HunyuanVideo

HunyuanVideo

HunyuanVideo 是一个可基于文本生成高质量图像和视频的项目。

HunyuanVideo 是一个专注于文本到图像及视频生成的项目。它具备强大的视频生成能力,支持多种分辨率和视频长度选择,能根据用户输入的文本生成逼真的图像和视频。使用先进的技术架构和算法,可灵活调整生成参数,满足不同场景的需求,是文本生成图像视频领域的优质工具。

WebUI for Browser Use

WebUI for Browser Use

一个基于 Gradio 构建的 WebUI,支持与浏览器智能体进行便捷交互。

WebUI for Browser Use 是一个强大的项目,它集成了多种大型语言模型,支持自定义浏览器使用,具备持久化浏览器会话等功能。用户可以通过简洁友好的界面轻松控制浏览器智能体完成各类任务,无论是数据提取、网页导航还是表单填写等操作都能高效实现,有利于提高工作效率和获取信息的便捷性。该项目适合开发者、研究人员以及需要自动化浏览器操作的人群使用,在 SEO 优化方面,其关键词涵盖浏览器使用、WebUI、大型语言模型集成等,有助于提高网页在搜索引擎中的曝光度。

xiaozhi-esp32

xiaozhi-esp32

基于 ESP32 的小智 AI 开发项目,支持多种网络连接与协议,实现语音交互等功能。

xiaozhi-esp32 是一个极具创新性的基于 ESP32 的开发项目,专注于人工智能语音交互领域。项目涵盖了丰富的功能,如网络连接、OTA 升级、设备激活等,同时支持多种语言。无论是开发爱好者还是专业开发者,都能借助该项目快速搭建起高效的 AI 语音交互系统,为智能设备开发提供强大助力。

olmocr

olmocr

一个用于 OCR 的项目,支持多种模型和服务器进行 PDF 到 Markdown 的转换,并提供测试和报告功能。

olmocr 是一个专注于光学字符识别(OCR)的 Python 项目,由 Allen Institute for Artificial Intelligence 开发。它支持多种模型和服务器,如 vllm、sglang、OpenAI 等,可将 PDF 文件的页面转换为 Markdown 格式。项目还提供了测试框架和 HTML 报告生成功能,方便用户对 OCR 结果进行评估和分析。适用于科研、文档处理等领域,有助于提高工作效率和准确性。

飞书多维表格

飞书多维表格

飞书多维表格 ×DeepSeek R1 满血版

飞书多维表格联合 DeepSeek R1 模型,提供 AI 自动化解决方案,支持批量写作、数据分析、跨模态处理等功能,适用于电商、短视频、影视创作等场景,提升企业生产力与创作效率。关键词:飞书多维表格、DeepSeek R1、AI 自动化、批量处理、企业协同工具。

下拉加载更多