The summer school will consist of tutorials to provide background knowledge, as well as invited talks on recent research works on agentic AI.
The topics for tutorials include (subject to change):
- Agentic Architectures
- Hands-on session on building an agent
- Memory and Knowledge in Agentic AI
- Planning, Reasoning, and Reinforcement Learning
- Multi-modal Agents
- Agentic AI for Software Engineering
We are also planning a brainstorming session on open problems in smaller groups (topics to depend on interest of participants, e.g. deep research agents, computer-use agents, software engineering agents, AI scientist agents, embodied AI agents, human value alignment, etc.)
It would be helpful for participants to have basic machine learning and coding skills.
Programme
Session slides, where available, are updated under the speakers’ names.
Day 1 – Wednesday, August 5
Time | Session | Title of Research Talk within Session | Speaker(s) |
08:30-09:00 | Breakfast |
|
|
09:00-09:45 | Opening Keynote | The Long Road to Agentic AI | |
09:45-10:15 | MSRA Introduction and Opportunities |
| Lily Sun, MSRA Jingwen Wang, MSRA |
10:15-10:30 | Group Photo and Short Break |
|
|
10:30-12:30 | Tutorial: Agentic Architectures |
| Yan Xia, MSRA |
12:30-13:30 | Lunch Break |
|
|
13:30-16:00 | Frontier Transformation in Research – Demo & Sharing on Microsoft Copilot & Agent Platform With afternoon tea served |
| Yong Keat Poh, Microsoft Singapore Jerry Hung, Microsoft Singapore Hieu Nguyen Nhu, GBB Microsoft IQ Jana Lee, Microsoft Singapore |
16:00-16:30 | Invited Talk | Can Language Models “Think” Internally as Human Brains? | Mingqing Xiao, MSRA Singapore |
| 18:00-20:00 | Dinner for Speakers |
|
|
Day 2 – Thursday, August 6
Time | Session | Title of Research Talk within Session | Speaker(s) |
08:30-09:00 | Breakfast |
|
|
09:00-10:00 | Invited Keynote | Some Ideas for Better AI Agency through Self-play and Self-improvement | Yee Whye Teh, Oxford |
10:00-10:30 | Invited Talk | Experiential Learning | Tianzhu Ye, MSRA |
10:30-12:30 | Tutorial: Memory and Knowledge in Agentic AI | Improving AI Capabilities through Agentic Memory | Baotong Lu, MSRA Yihao Ang, Yixuan Tang, NUS |
12:30-13:30 | Lunch Break |
|
|
13:30-16:00 | Tutorial: Planning, Reasoning, and Reinforcement Learning With afternoon tea served | RL over Commodity Networks: Overcoming the Bandwidth Barrier with Lossless Sparse Deltas (Dehui Wei) | |
16:00-16:30 | Invited Talk | Forecast, Act, Adapt: Agentic Time-Series Modeling and Temporal Perspectives on AI Agent Design | Jiawen Zhang, MSRA Singapore |
Day 3 – Friday, August 7
Time | Session | Title of Research Talk within Session | Speaker(s) |
08:30-09:00 | Breakfast |
|
|
09:00-10:00 | Invited Keynote | Building Physical Agents via Structured Representations | |
10:00-12:00 | Tutorial: From Multi-modal Understanding to Multi-modal Agents |
| |
12:00-13:00 | Lunch Break |
|
|
13:00-14:00 | Tutorial: Agentic AI for Software Engineering |
| Abhik Roychoudhury, NUS |
14:00-14:30 | Invited Talk | Efficient Next-Generation AI Networking: From Scalable Fabrics to Fast Simulation Platforms | Ran Shu, MSRA |
14:30-16:30 | Open Problems Discussion With afternoon tea served |
| Xinxing Xu, MSRA Singapore (with other facilitators to be confirmed) |
Keynote Speakers
Prof Wee Sun Lee
NUS Deputy Director,
NUS Artificial Intelligence Institute (NAII)
Prof Yee Whye Teh
University of Oxford
Prof Jiajun Wu
Stanford University
Meet Our Speakers
Venue
Lecture Theatre 15 (LT15), AS6, Level 1,
11 Computing Drive, Singapore 117416
Participation Fee and Registration
Faculty, students and industry practitioners are welcome to participate in the Summer School.
Prerequisite: It would be helpful for participants to have basic machine learning and coding skills.
Participation Fee:
Faculty/Students - SGD $100 per participant (before GST and any prevailing credit card fees)
Industry - SGD $500 per participant (before GST and any prevailing credit card fees)
Registration Link:
Please click on your participant category below for the registration link.
You will be prompted for payment after you complete registration.
