Advanced Tool Learning and Selection System (ATLASS): A Closed-Loop Framework Using LLM

2025-03-13

AI Summary

Existing tool-augmented LLM systems rely on manually predefined tool sets, limiting their ability to handle complex, out-of-distribution tasks beyond their parametric knowledge. Method: We propose the first three-stage closed-loop tool learning framework—comprising requirement understanding, tool generation/retrieval, and task solving—that dynamically generates, retrieves, executes, and refines external tools. It automatically constructs Python tools from API documentation, enables environment self-configuration via online crawling and parsing, and incorporates human-in-the-loop safety verification for code reliability. The system integrates GPT-4, a Python interpreter, and persistent tool storage to transcend static tool-set constraints. Contribution/Results: Experiments demonstrate significant improvements in LLM success rates on cross-domain API and package invocation tasks. Generated tools are reusable, incur low inference overhead, and exhibit strong generalization and execution reliability.

Technology Category

• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Machine Learning: Large Multimodal Models (LMMs)
• Humans and AI: Human-in-the-loop Machine Learning

Application Category

• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

The combination of LLM agents with external tools enables models to solve complex tasks beyond their knowledge base. Human-designed tools are inflexible and restricted to solutions within the scope of pre-existing tools created by experts. To address this problem, we propose ATLASS, an advanced tool learning and selection system designed as a closed-loop framework. It enables the LLM to solve problems by dynamically generating external tools on demand. In this framework, agents play a crucial role in orchestrating tool selection, execution, and refinement, ensuring adaptive problem-solving capabilities. The operation of ATLASS follows three phases: The first phase, Understanding Tool Requirements, involves the Agents determining whether tools are required and specifying their functionality; the second phase, Tool Retrieval/Generation, involves the Agents retrieving or generating tools based on their availability; and the third phase, Task Solving, involves combining all the component tools necessary to complete the initial task. The Tool Dataset stores the generated tools, ensuring reusability and minimizing inference cost. Current LLM-based tool generation systems have difficulty creating complex tools that need APIs or external packages. In ATLASS, we solve the problem by automatically setting up the environment, fetching relevant API documentation online, and using a Python interpreter to create a reliable, versatile tool that works in a wider range of situations. OpenAI GPT-4.0 is used as the LLM agent, and safety and ethical concerns are handled through human feedback before executing generated code. By addressing the limitations of predefined toolsets and enhancing adaptability, ATLASS serves as a real-world solution that empowers users with dynamically generated tools for complex problem-solving.

Problem

• Dynamic tool generation for complex tasks using LLM.
• Overcoming inflexibility of human-designed, predefined tools.
• Enhancing adaptability and reusability in problem-solving frameworks.

Innovation

• Dynamic tool generation using LLM agents
• Closed-loop framework for adaptive problem-solving
• Automatic environment setup and API integration