Advanced Topics in Statistics (461-0-20)

Topic

Agentic AI

Instructors

Kaize Ding

Meeting Info

Lunt Hall 105: Mon, Wed 5:00PM - 6:20PM

Overview of class

Topic: Agentic AI
This course explores modern approaches to building autonomous AI agents and agentic systems. Students will learn how to design AI systems that can actively interact with environments, make decisions, and improve through experience. Topics include agent architectures, planning and reasoning, reinforcement learning, multi-agent systems, tool use and agent memory, etc. The course combines conceptual topics with hands-on projects implementing autonomous agents for real-world applications.

Registration Requirements

This course requires a background and knowledge in machine learning and large language models. Permission number required from instructor.

Learning Objectives

Understand state-of-the-art approaches in agent-based AI systems
Design and implement autonomous AI agents with decision-making capabilities
Evaluate and debug agentic AI systems using appropriate metrics and methodologies

Teaching Method

Combination of lectures, paper discussions, and project assignments. Students will work on a final project implementing an autonomous agent system.

Evaluation Method

Individual Presentations and Assignments: 20%
Group Presentations and Discussions: 20%
Project: 50%
Class Participation: 10%

Class Materials (Required)

No required textbook. All readings will be provided from recent research papers and will be available through Canvas.