JARVIS: A Multi-Agent Code Assistant for High-Quality EDA Script Generation

2025-05-20

AI Summary

This work addresses the dual challenges of data scarcity in electronic design automation (EDA) and hallucination in large language models (LLMs), which severely degrade the quality of generated EDA scripts. To this end, we propose the first multi-agent code assistant framework specifically designed for EDA. Our framework integrates a domain-adapted LLM—fine-tuned on synthetically generated EDA data—with a structured compiler, a rule-driven script repair module, and an enhanced retrieval mechanism, enabling joint semantic and syntactic validation to substantially mitigate hallucination. Evaluated across multiple EDA benchmarks, our approach achieves state-of-the-art accuracy and reliability, outperforming existing methods across all metrics. It marks the first demonstration of high-fidelity, verifiable, and domain-specific EDA script generation, establishing a novel paradigm for deploying LLMs in precision-critical engineering applications.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Multiagent Systems: Adversarial Agents

Application Category

• 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
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

This paper presents JARVIS, a novel multi-agent framework that leverages Large Language Models (LLMs) and domain expertise to generate high-quality scripts for specialized Electronic Design Automation (EDA) tasks. By combining a domain-specific LLM trained with synthetically generated data, a custom compiler for structural verification, rule enforcement, code fixing capabilities, and advanced retrieval mechanisms, our approach achieves significant improvements over state-of-the-art domain-specific models. Our framework addresses the challenges of data scarcity and hallucination errors in LLMs, demonstrating the potential of LLMs in specialized engineering domains. We evaluate our framework on multiple benchmarks and show that it outperforms existing models in terms of accuracy and reliability. Our work sets a new precedent for the application of LLMs in EDA and paves the way for future innovations in this field.

Problem

• Generates high-quality EDA scripts using multi-agent LLMs
• Addresses data scarcity and hallucination errors in LLMs
• Improves accuracy and reliability in specialized engineering domains

Innovation

• Multi-agent framework with domain-specific LLM
• Custom compiler for verification and rule enforcement
• Advanced retrieval mechanisms to reduce errors