Skip to content

series2.md

Latest commit

 

History

History
184 lines (157 loc) · 5.13 KB

series2.md

File metadata and controls

184 lines (157 loc) · 5.13 KB
title Series Ⅱ
layout default
nav_order 4

Series Ⅱ

<style> .staff-grid { display: grid; grid-template-columns: repeat(2, 1fr); /* 固定每行两列 */ gap: 20px; /* 卡片间距 */ justify-items: center; /* 卡片居中 */ } .staff-card{ background:#fff; padding:12px; border-radius:12px; box-shadow:0 1px 2px rgba(0,0,0,.06); } .staff-card img{ width:100%; height:160px; object-fit:cover; border-radius:8px; display:block; } .staff-card p{ margin:8px 0 0; } .staff-card a{ font-weight:600; } </style>

Lecture Guests

Sabina (Yudi) Nong

Sabina (Yudi) Nong
Master of International Policy @ Stanford FSI
sabn@stanford.edu

Kumar Shridhar

Kumar Shridhar
PhD @ ETH Zürich in Machine Learning
shridhar.stark@gmail.com

Yuan Tian

Yuan Tian
Postdoc @ ETH Zurich

Bang Liu

Bang Liu
Associate Professor @ University of Montreal (UdeM)

Schedule

Date Guest Lecture Supplemental Readings
August 29th Governance: LLM + AI Agents
Sabina (Yudi) Nong, Stanford

<strong>Key Points:</strong><br>
<ul>
  <li>AI Governance & its Stakes</li>
  <li>Why AI Agents Require a Distinct Governance Lens (Procedural vs. Predictive)</li>
  <li>Developer Governance</li>
  <li>Regulatory Governance</li>
  <li>What does a good governance structure look like?</li>
</ul>

<a href="https://docs.google.com/presentation/d/145F-UnXUBswnz03WvZkliK7eLiLyhGY0bTrDJo8VUAk/mobilepresent?slide=id.g376427fc988_0_361" target="_blank">Slides (Website)</a> · 
<a href="https://www.youtube.com/watch?v=MpbzMipStXk&pp=0gcJCU0KAYcqIYzv" target="_blank">Recording</a>
To be updated
September 3rd Interactive symbolic regression with co-design mechanism
Yuan Tian, ETH Zurich

<strong>Key Topics:</strong><br>
<ul>
  <li>Introduction to symbolic regression and its applications in scientific discovery</li>
  <li>Challenges in traditional symbolic regression approaches</li>
  <li>Co-design mechanism: integrating human expertise with machine learning</li>
  <li>Interactive frameworks for collaborative model discovery</li>
  <li>Case studies and practical applications in scientific computing</li>
</ul>

<a href="https://www.youtube.com/watch?v=Q-l6pVR3z-Y" target="_blank">Recording</a>
To be updated
October 3rd Rewiring the reward pathways
Kumar Shridhar, ETH Zürich

<strong>Key Topics:</strong><br>
<ul>
  <li>Reward models in ML pipeline</li>
  <li>Brittleness of current reward models</li>
  <li>What to do when there is no clear reward (non verifiable rewards)</li>
  <li>How to rewire reward pathways</li>
  <li>A common reward model for both verifiable and non verifiable rewards</li>
</ul>

<a href="https://www.youtube.com/watch?v=x7wvhD28EMU" target="_blank">Recording</a>
To be updated
October 17th Advances And Challenges In Foundation Agents - Building The Cognitive Engine
Bang Liu, University of Montreal (UdeM)

<strong>Key Topics:</strong><br>
<ul>
  <li>Foundation agents: architecture and core components</li>
  <li>Building cognitive engines for autonomous decision-making</li>
  <li>Challenges in scaling foundation agents to complex environments</li>
  <li>Integration of perception, reasoning, and action in agent systems</li>
  <li>Current limitations and future directions in foundation agent research</li>
</ul>

<a href="https://www.youtube.com/watch?v=tPfnFYgFNig" target="_blank">Recording</a>
To be updated