Topic 12: Agentic AI for Science

Learning Goals

Be able to:

• Articulate why scientific research is inherently agentic and why the agent loop (plan → act → observe → revise) maps naturally onto the scientific method

• Describe the emerging "scientific agent stack" and connect it to architectures they have already built (ReAct, tool calling, MCP, RAG, A2A)

• Identify concrete agentic AI systems across the biological sciences (drug discovery, protein engineering) and physical sciences (materials discovery, autonomous laboratories, chemistry)

• Explain how the Ai2 tool ecosystem (Semantic Scholar, OpenScholar, ScholarQA, Asta) provides the knowledge infrastructure that scientific agents depend on

• Analyze the JFC framework (Moreno et al., 2026) as an integrative case study that combines autonomous agent execution, literature RAG, and multi-agent review to perform experimental high-energy physics

• Critically evaluate the challenges of deploying agents in high-stakes scientific domains: hallucination, verification, reproducibility, and the limits of autonomy

Tasks

Work on your final project

Lesson Plan and Slides

Agentic AI for science lesson plan

Agentic AI for science slides

Resources

• Agarwal, Dhruv et al (2025), AutoDiscovery: Open-ended Scientific Discovery via Bayesian Surprise, NeurIPS 2025, https://arxiv.org/abs/2507.00310

• Eric A. Moreno, Samuel Bright-Thonney, Andrzej Novak, Dolores Garcia, Philip Harris (2026). AI Agents Can Already Autonomously Perform Experimental High Energy Physics. https://inspirehep.net/literature/3132489

• HEPTAPOD: (High Energy Physics (HEP) Toolkit for Agentic Planning, Orchestration and Deployment) is a publicly available AI agent orchestration framework for particle physics simulation and analysis workflows built on the Orchestral AI Framework. HEPTAPOD Paper: [https://arxiv.org/abs/2512.15867]

• Github Repo: https://github.com/tonymenzo/heptapod

• Orchestral AI: https://orchestral-ai.com