CSCE 631-D: Intelligent Agents

Computational Game Solving

Fall 2026 · Alan Kuhnle · Texas A&M University

Course overview

Intelligent agents must reason about the choices of other agents. This course develops the game-theoretic and computational foundations for doing so, then connects them to modern LLM agents, multi-agent interaction, and AI monitoring. Students will implement and empirically evaluate agents using the TAMU AI API.

Meeting: Tuesday and Thursday, 9:35–10:50 AM, Zach 592 (in-person), with a web section. Office hours: PETR 421; remote via Zoom.

Canvas: in-person section Canvas: web section In-person syllabus (PDF) Online syllabus (PDF) TAMU API Guide (PDF)

Module overview

Weeks 1–15 · 75-minute lectures, except the M0 orientation

Week 1 · 30-minute orientation

M0: Course Overview

Course expectations, Canvas, TAMU API setup, integrity, and accessibility.

View module
Weeks 1–2 · four lectures

M1: Foundations of Strategic Games

Normal-form games, search, equilibria, and learning.

View module
Weeks 3–4 · four lectures

M2: Algorithms for Equilibria

Complexity, exact methods, and approximation.

View module
Week 5 · two lectures

M3: Regret Minimization

Extensive-form and repeated games.

View module
Week 6 · two lectures

M4: Extensive-Form Games

Information sets and sequential rationality.

View module
Week 7 · two lectures

M5: CFR and Self-Play

Regret minimization and counterfactual regret.

View module
Week 8 · two lectures

M6: Abstraction and Agent Action Spaces

Agent action spaces and evaluation.

View module
Week 9 · two lectures

M7: Multi-Agent LLM Systems

LLM game players, communication, and coordination.

View module
Week 10 · two lectures

M8: Monitoring Autonomous Agents

Monitoring, alignment, and submodular optimization.

View module
Weeks 11–12 · two lectures + project time

M9: Applications, Case Studies, and Frontiers

Mechanism design, fairness, and project presentations.

View module