AI Co-Learning

Large Language Models (LLMs) have profoundly reshaped the landscape of artificial intelligence. This course is designed to guide learners through the core principles of LLMs and intelligent agents, combining theoretical foundations with hands-on experience.

We will explore:

  • The architecture and capabilities of large language models
  • The key features needed to automate real-world tasks
  • The underlying framework for building and deploying agent systems

In addition to foundational knowledge, we will delve into practical applications, including:

  • Code generation
  • Robotic control
  • Web process automation

We will also examine the limitations and risks of current LLM agents — such as hallucination, misalignment, and brittleness — and discuss how future systems might address these challenges.

Topics Covered

🔹 Foundation of LLMs
🔹 LLM Inference & Reasoning
🔹 Agent Overview (Planning – Tool Use – Memory)
🔹 MCP (Modular Control Protocol) & A2A (Agent-to-Agent communication)

Lecture Guests

Michael Mahoney

Michael Mahoney
Professor in University of California Berkeley, Amazon Scholar
mmahoney@stat.berkeley.edu

Zhou Zijian

Zhou Zijian
PhD in Computer Science at National University of Singapore
zhou_zijian@u.nus.edu

Yuxi Li

Yuxi Li
PhD in Computer Science at University of Alberta
yuxili@gmail.com

Zhi Wang

Zhi Wang
PhD in Machine Learning at University of California San Diego
zhw119@ucsd.edu

Schedule

Date Guest Lecture Supplemental Readings
July 18th LLM Inference and Reasoning
Zhou Zijian, NUS

Inference:
- What are the inputs and outputs of an LLM model?
- Difference between pre-filling and auto-regressive decoding
- Auto-regressive decoding:
    - How are tokens sampled based on output
    - What are top-k, top-p, temperature?
    - How does the LLM know when to stop?

Reasoning:
- What is reasoning in its fundamental sense?
- Why reasoning is important for LLM?
- Two approaches of achieving reasoning:
    - Using a fine-tuned model
    - Prompting
Slides Recording 		- Chain-of-Thought Reasoning Without Prompting
- Large Language Models as Optimizers
July 23rd Post-Training Reasoning Models
Zhi Wang, UCSD

Recording / Slides:
To be uploaded 		- Introduction to Reinforcement Learning
- Introduction to Deep Learning
August 1rd Learning from Experience AKA Reinforcement Learning (2024 Turing Award topic for research and business)
Yuxi Li, University of Alberta, AI4All Institute

First principles:
- Learning from experience
- Iterative improvement based on ground truth
Research:
- Pursuing truth or following trend
- Autonomous, optimal and adaptive agents
- Simulation, integration of (world) model and data
- Explore alternative approaches w.r.t. data collection, architectures, and algorithms
- "Small" language models, modularity, generalist vs specialist
Business:
- Value investment
- AI vs IT
- Code LLMs
- Experience data collection
- Decentralized AI aka AI + blockchain, in particular, for stablecoin
Recording / Slides:
To be uploaded 		To be uploaded

Supported by SpoonOS
Follow on X (Twitter)