NeurIPS 2025 Workshop Schedule - PokéAgent Challenge

Workshop Schedule

8:00 – 8:30 AM

The PokeAgent Challenge: Competitive and Long Context Learning at Scale

Seth Karten (Princeton University)
Opening presentation introducing the PokeAgent Challenge, its scientific relevance and challenges, and key research frontiers in game AI.

8:30 – 9:00 AM

Invited Talk 1: Pokémon as an AI Problem

Aaron Traylor (Work conducted independently)
A deep dive into strategic decision-making under imperfect information and their applications in AI systems from Pokémon VGC (Video Game Championship) World Finalist.

9:00 – 9:30 AM

Invited Talk 2: Gemini Game Effort and Kaggle Game Arena

Minmin Chen (Google DeepMind)
Exploring Google DeepMind's game AI initiatives, including the Gemini project and Kaggle's Game Arena platform.

9:30 – 10:00 AM

Winner Awards + Lightning Presentations

Top Teams (5 minutes each)
Celebration and recognition of winning teams with brief presentations showcasing their innovative approaches and solutions.

10:00 – 10:30 AM

Panel Discussion: Research Frontiers in Game AI

Moderators: Jake Grigsby (UT Austin) & Seth Karten (Princeton)
Panelists: Stephanie Milani (NYU / Johns Hopkins), Fei Fang (CMU), Chi Jin (Princeton), Minmin Chen (Google DeepMind), Aaron Traylor (Work conducted independently)
An engaging discussion on the future of game AI research, specialist RL agents vs generalist LLM agents, emerging challenges, and opportunities for advancing foundation agents.

10:30 – 10:45 AM

Closing Remarks

Chi Jin (Princeton University)
Final thoughts and reflections on the competition, with a look ahead to future directions in learning foundation agents from games.