JARVIS: A Multi-Agent Code Assistant for High-Quality EDA Script Generation
This work addresses data scarcity and hallucination errors in LLMs for specialized engineering domains like EDA, representing an incremental advancement in applying LLMs to niche tasks.
The paper tackles the problem of generating high-quality Electronic Design Automation (EDA) scripts by introducing JARVIS, a multi-agent framework that combines domain-specific LLMs, verification, and retrieval mechanisms, achieving significant improvements in accuracy and reliability over state-of-the-art models on benchmarks.
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.