Enhancing Agentic Autonomous Scientific Discovery with Vision-Language Model Capabilities

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Gandhi, Kahaan, Bolliet, Boris, Zubeldia, Inigo
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866912716623446016
author Gandhi, Kahaan
Bolliet, Boris
Zubeldia, Inigo
author_facet Gandhi, Kahaan
Bolliet, Boris
Zubeldia, Inigo
contents 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
format Preprint
id arxiv_https___arxiv_org_abs_2511_14631
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Enhancing Agentic Autonomous Scientific Discovery with Vision-Language Model Capabilities
Gandhi, Kahaan
Bolliet, Boris
Zubeldia, Inigo
Computation and Language
Artificial Intelligence
Computer Vision and Pattern Recognition
Multiagent Systems
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
title Enhancing Agentic Autonomous Scientific Discovery with Vision-Language Model Capabilities
topic Computation and Language
Artificial Intelligence
Computer Vision and Pattern Recognition
Multiagent Systems
url https://arxiv.org/abs/2511.14631