SPARC: A Multi-Agent System for Electrical Circuit Question Answering
It addresses the challenge of complex mathematical reasoning in multimodal LLMs for circuit diagram QA, a domain-specific problem.
SPARC is a multi-agent system for answering questions about electrical circuit diagrams that uses LLM agents to create and run physics-based simulations, achieving 83% accuracy and up to 58% absolute improvement over baselines.
Electrical circuit diagram QA tasks require complex mathematical reasoning, which remains challenging for multimodal LLMs. We present SPARC, a multi-agent system that answers questions over circuit diagrams by grounding reasoning in executable physics-based simulations. SPARC uses LLM agents to synthesize, execute, and analyze simulation programs, improving accuracy and reliability by design. It achieves 83% accuracy, with up to a 58% absolute improvement over baselines, while enabling systematic error diagnosis.