Stanford
University
  • Stanford Home
  • Maps & Directions
  • Search Stanford
  • Emergency Info
  • Terms of Use
  • Privacy
  • Copyright
  • Trademarks
  • Non-Discrimination
  • Accessibility
© Stanford University.  Stanford, California 94305.
Wolfgang Lehrach | Code World Models for General Game Playing | Stanford HAI

Stay Up To Date

Get the latest news, advances in research, policy work, and education program updates from HAI in your inbox weekly.

Sign Up For Latest News

Navigate
  • About
  • Events
  • AI Glossary
  • Careers
  • Search
Participate
  • Get Involved
  • Support HAI
  • Contact Us
Skip to content
  • About

    • About
    • People
    • Get Involved with HAI
    • Support HAI
    • Subscribe to Email
  • Research

    • Research
    • Research Programs
    • Grants
    • Marlowe (opens in new tab)
    • Student Affinity Groups
    • Centers & Labs
    • Research Publications
    • Research Partners
  • Education

    • Education
    • Executive and Professional Education
    • Government and Policymakers
    • K-12
    • Stanford Students
  • Policy

    • Policy
    • Policy Publications
    • Policymaker Education
    • Student Opportunities
  • AI Index

    • AI Index
    • AI Index Report
    • Global Vibrancy Tool
    • People
  • News
  • Events
  • Industry
  • Centers & Labs
Your browser does not support the video tag.
eventSeminar

Wolfgang Lehrach | Code World Models for General Game Playing

Status
Past
Date
Wednesday, May 13, 2026 12:00 PM - 1:15 PM PST/PDT
Location
353 Jane Stanford Way, Stanford, CA, 94305 | Room 119
Topics
Machine Learning
Natural Language Processing
Attend Virtually
Overview
Watch event recording

While Large Language Models (LLMs) show promise in many domains, relying on them for direct policy generation in games often results in illegal moves and poor strategic play.

Overview
Watch event recording
Share
Link copied to clipboard!
Event Contact
Stanford HAI
stanford-hai@stanford.edu

Related Events

Rob Reich, Diyi Yang, Amy Zegart | Prompt Response: Should AI Slow Down?
WebinarSep 21, 202610:00 AM - 11:15 AM
September
21
2026

HAI faculty offer independent critique of the AI debates of the moment.

Webinar
Your browser does not support the video tag.

Rob Reich, Diyi Yang, Amy Zegart | Prompt Response: Should AI Slow Down?

Sep 21, 202610:00 AM - 11:15 AM

HAI faculty offer independent critique of the AI debates of the moment.

Alexandr Lenk & Arvind Karunakaran | Industry Conversation with Instacart
SeminarSep 23, 202612:00 PM - 1:15 PM
September
23
2026

This seminar pairs a case study of AI diffusion with an organizational approach to technological changes in the workplace.

Seminar
Your browser does not support the video tag.

Alexandr Lenk & Arvind Karunakaran | Industry Conversation with Instacart

Sep 23, 202612:00 PM - 1:15 PM

This seminar pairs a case study of AI diffusion with an organizational approach to technological changes in the workplace.

Marlowe | AI + Data for Science with Stephen Baccus
SeminarSep 23, 2026
September
23
2026

Sessions run Wednesdays from 4:30–5:30 PM in CoDa E160. Each session features a different Stanford speaker; talk titles are announced by the organizers.

Event

Marlowe | AI + Data for Science with Stephen Baccus

Sep 23, 2026

Sessions run Wednesdays from 4:30–5:30 PM in CoDa E160. Each session features a different Stanford speaker; talk titles are announced by the organizers.

In this talk, I present an approach that moves away from direct prompting, instead using LLMs as program synthesizers to bridge the gap between natural language rules and symbolic world models.  The LLM receives a game description and example trajectories, and outputs an executable, symbolic world model (CWM) represented in Python. The trajectories also ensure the rules are correctly captured and aid in refining the CWM if they are not.  Note that even trajectories containing only a single player's observations and actions can be used to help validate and refine CWMs.  Furthermore, partially observed trajectories also allow comparisons between CWMs via a bound on the likelihood.  


Given a CWM, Monte Carlo Tree Search (MCTS) or Reinforcement Learning (RL) methods can play the game, and gameplay can be further enhanced by adding in LLM-derived synthesized value functions. Imperfect information games are handled by having the LLM synthesize inference functions to impute information sets, or by directly training reinforcement learning policies on top of the CWM.

Speaker
Wolfgang Lehrach
Staff Research Scientist, DeepMind