Training through self-play
We train policies through repeated interaction and competition, using feedback to improve decisions over successive rounds of training.
Reinforcement learning · Adaptive agents
Intelligence, learned
through experience.
We build learning systems that turn interaction into capability. Our work combines reinforcement learning, memory, and self-play to train agents that improve through experience.
From training to competition
First place at generals.bot
We trained and deployed a self-play agent that took first place in a generals.bot competition. The system combines recurrent reinforcement learning, a custom self-play pipeline, and an optimized CPU inference implementation.
Decisions with memory
Our recurrent agent uses memory to make decisions from partial observations. This is a starting point for studying behavior over longer horizons.
Beyond the first task
We aim to make these learning methods reusable across tasks, with language agents and robotics as longer-term directions.