The mission of the EMERGE lab is to make it easy and efficient to develop capable, safe, and intelligent multi-agent systems through learning, simulation, and data. We emphasise use-inspired basic research, turning serious efforts to solve problems in autonomy, systems, and transportation into new algorithms, insights, and discoveries. If there's a unifying thread to our research, it's that the fastest way to understand something is to start building.
In practice, our tools consist heavily of scaled reinforcement learning, fast and diverse simulation, human data, and solid engineering. A smattering of interests right now:
- Scaling up search in multi-agent RL
- Getting meta-reinforcement learning working
- Designing accurate reactive models of human behavior
- Building an open-source planning stack for a self-driving car
News
- 📝 [Jun. 2026] With collaborators, we show that self-play can be scaled to end-to-end driving without any human data via fast simulators — Scaling Self-Play for End-to-End Driving.
- 📝 [Jun. 2026] New paper: just 30 minutes of human data is enough to build capable, but human-like agents — Human-like autonomy emerges from self-play and a pinch of human data.
- 🚀 [May 2026] Congrats to Daphne and Julian on their NVIDIA internships, and Aditya on his Amazon internship!
- 🤝 [Jan. 2026] We've started a new collaboration with Torc deploying self-play RL to autonomous trucking.
- 📝 [Dec. 2025] We're open-sourcing PufferDrive 2.0, a driving simulator that does end-to-end RL training at 300k SPS. See more at the blog post or watch the video.
- 🤝 [Dec. 2025] Excited to announce a new collaboration with Motional to develop the next generation of driving benchmarks via procedurally generated adversarial environments.
- 🤝 [Sept. 2025] Excited to announce a new collaboration with Qualcomm to deploy RL-based driving agents. We are hiring PhD students and postdocs for this project.
- 📝 [May 2025] We've built and open-sourced highly reliable RL-based driving agents! Read the paper.
- 🚀 [May 2025] Daphne Cornelisse is off to Waymo for the summer and Kevin Joseph to Applied Intuition! Congrats folks!
- 📝 [Feb. 2025] New scalable benchmarks support the hypothesis that self-play policy-gradient methods excel at two-player zero-sum games. Paper · Play our agents.