Call for Papers
Recent breakthroughs in Large Language Models (LLMs) and foundation models have catalyzed transformative applications across scientific domains, presenting unprecedented opportunities for accelerating research and discovery. These AI systems show increasing promise in supporting the full scientific research lifecycle—from idea generation and literature synthesis to experimental design, data analysis, and hypothesis validation. Despite their rapidly evolving capabilities, foundation models still face significant challenges in scientific contexts, including reliability, factuality, attribution, and ethical use. The expanding intersection between AI and science necessitates a collaborative approach that brings together expertise from many communities.
This workshop aims to bridge the gap between AI researchers and domain scientists by fostering interdisciplinary dialogue on how foundation models can enhance scientific reasoning, assist human researchers, and transform scientific workflows. We welcome contributions exploring both generalist models (e.g., GPT, Claude, Gemini) and specialist models trained on scientific data, with particular emphasis on real-world applications that demonstrate scientific reasoning, assistance capabilities, and novel collaboration paradigms. By bringing together diverse perspectives, we hope to catalyze research that moves beyond using AI as merely an information retrieval tool toward systems that can meaningfully participate in and accelerate the scientific process itself.
Through our interdisciplinary program, we aim to foster new connections for this emerging community.
Topics of Interest
We welcome submissions on topics including but not limited to:
Scientific Reasoning and Capabilities:
- Domain-specific foundation models for scientific applications
- Quantitative, symbolic, and physics-based reasoning in foundation models
- High-precision scientific tool use (e.g., statistical analysis, laboratory automation)
- Modeling structured scientific information (equations, code, plots, molecular graphs)
- Compositional generalization and abstraction in scientific domains
AI Systems for Scientific Assistance:
- End-to-end research assistants and agents for scientific workflows
- Literature search, synthesis, and question-answering across scientific domains
- Multimodal scientific document understanding and knowledge extraction
- Experiment design, execution, and iteration with AI assistance
- Novel human-AI collaboration paradigms in scientific research
Evaluation, Safety, and Ethics:
- Datasets, benchmarks, and evaluation methodologies for scientific AI systems
- Attribution, factuality, and hallucination mitigation in scientific contexts
- Safety guardrails for autonomous scientific exploration
- Ethical considerations in AI-assisted research, including dual-use concerns
- Perspectives on the evolving role of AI in scientific discovery
Submissions will be non-arxival, with an 8-page limit.
Deadlines
- Submission deadline: June 23, 2025 11:59pm AOE
- Decision Notification: July 24, 2025 AOE