CS 839 · Fall 2026
AI Agents for Science
Meeting time. Tuesdays and Thursdays, 9:30–10:45 AM, in Mechanical Engineering 1164. Readings, presentation assignments, and project deadlines will be added before the semester.
| Date | Session | Focus |
|---|---|---|
| Part I. Foundations and capabilities of LLM Agents | ||
| Thu, Sept. 3 | Lecture 1: Introduction: AI Agents for Science | Overview. Agents, LLMs, and scientific discovery |
| Tue, Sept. 8 | Lecture 2: From Language Models to Agents | Instruction following, reasoning models, and agent loops |
| Thu, Sept. 10 | Lecture 3: Agent Architectures and Frameworks | Orchestration, state, control flow, debugging |
| Tue, Sept. 15 | Lecture 4: Tools, Environments, and Agent–Computer Interfaces | Function calling, code execution, scientific software, and error recovery |
| Thu, Sept. 17 | Lecture 4 (continued): Tools, Environments, and Agent–Computer Interfaces | Continued from Sept. 15 |
| Tue, Sept. 22 | Lecture 5: Planning and Task Decomposition | ReAct, search, reflection, verification, and long-horizon tasks |
| Part II. Scientific applications | ||
| Thu, Sept. 24 | Lecture 7: Agents for Scientific Literature and Evidence Synthesis | Search, citation verification, systematic reviews, and research gaps |
| Tue, Sept. 29 | Lecture 8: Agents for Scientific Coding, Data Analysis, and Simulation | Analysis code, simulators, workflow debugging, and reproducibility |
| Thu, Oct. 1 | Lecture 9: Agents for Hypothesis Generation and Experimental Design | Hypotheses, experiment selection, active learning, and uncertainty |
| Tue, Oct. 6 | Lecture 10: Auto-Laboratories and Autonomous Experimentation | Robotics, closed-loop experimentation, costly actions, and approval |
| Part III. Evaluation, learning, and open frontiers | ||
| Thu, Oct. 8 | Lecture 11: Evaluation and Benchmarks for Scientific Agents | Scientific validity, baselines, reproducibility, cost, and discovery |
| Tue, Oct. 13 | Lecture 12: Training and Improving Agents | Trajectory data, process supervision, reinforcement learning, and credit assignment |
| Thu, Oct. 15 | Lecture 13: Multi-Agent Scientific Discovery | Specialized roles, debate, peer review, coordination, and failure |
| Tue, Oct. 20 | Lecture 14: Human–Agent Interaction, Safety, and Scientific Integrity | Oversight, sandboxing, credentials, evidence, attribution, and novelty |
| Thu, Oct. 22 | Lecture 15: Scientific Case Studies and Open Frontiers | Biology, chemistry, materials, mathematics, climate, and new applications |
| Part IV. Student presentations | ||
| Tue, Oct. 27 | Presentation and Project Workshop | Paper selection, expectations, and presentation planning |
| Thu, Oct. 29 | Student Presentation 1 | TBA |
| Tue, Nov. 3 | Student Presentation 2 | TBA |
| Thu, Nov. 5 | Student Presentation 3 | TBA |
| Tue, Nov. 10 | Student Presentation 4 | TBA |
| Thu, Nov. 12 | Student Presentation 5 | TBA |
| Tue, Nov. 17 | Student Presentation 6 | TBA |
| Thu, Nov. 19 | Student Presentation 7 | TBA |
| Tue, Nov. 24 | Student Presentation 8 | TBA |
| Thu, Nov. 26 | No class | Thanksgiving recess |
| Tue, Dec. 1 | Student Presentation 9 | TBA |
| Thu, Dec. 3 | Student Presentation 10 | TBA |
| Tue, Dec. 8 | Course Synthesis and Project Discussion | What have we learned, and what is next? |