Enhancing Agentic Autonomous Scientific Discovery with Vision-Language Model Capabilities
This addresses the challenge of improving end-to-end autonomous scientific discovery for researchers in fields like cosmology and astrochemistry, representing a novel method for a known bottleneck.
The paper tackles the problem of autonomous scientific discovery by using vision-language models to guide multi-agent systems, achieving pass at 1 scores of 0.7-0.8 on a 10-task benchmark compared to 0.2-0.3 for code-only baselines.
We show that multi-agent systems guided by vision-language models (VLMs) improve end-to-end autonomous scientific discovery. By treating plots as verifiable checkpoints, a VLM-as-a-judge evaluates figures against dynamically generated domain-specific rubrics, enabling agents to correct their own errors and steer exploratory data analysis in real-time. Case studies in cosmology and astrochemistry demonstrate recovery from faulty reasoning paths and adaptation to new datasets without human intervention. On a 10-task benchmark for data-driven discovery, VLM-augmented systems achieve pass at 1 scores of 0.7-0.8, compared to 0.2-0.3 for code-only and 0.4-0.5 for code-and-text baselines, while also providing auditable reasoning traces that improve interpretability. Code available here: https://github.com/CMBAgents/cmbagent