Excited to announce playground.mujoco.org an open-source library for GPU-accelerated robot learning and sim-to-real transfer. With `pip install playground` you get access to a whole host of robotic learning environments, many of which were transferred onto robots!

Had the delight of contributing during my internship at #GoogleDeepMind, collaborating with incredible collaborators from Berkeley, U. Toronto, Cambridge, Stanford and @deepmind.google.web.brid.gy.

Learn more here: playground.mujoco.org

We are so excited to share MuJoCo Playground—an open-source sim-to-real pipeline harnessing GPU-accelerated reinforcement learning and environment simulation with JAX, tackling complex real-world tasks with training in as little as 10 minutes!

It was a delight to contribute during my internship at
#GoogleDeepMind, collaborating with incredible folks from
from Berkeley, U. Toronto, Cambridge, Stanford and
@deepmind.google.web.brid.gy.

Dive in here:
playground.mujoco.org

We’re also excited to share MuJoCo Playground itself—a broader open-source sim-to-real pipeline harnessing GPU-accelerated reinforcement learning and environment simulation with JAX, tackling complex real-world tasks with training in as little as 10 minutes.

Thanks for the shout-out, @mwulfmeier.bsky.social! Thrilled to release our block reorientation setup from my #GoogleDeepMind internship, now in MuJoCo Playground. Leveraging domain randomization and curriculum learning, we achieve smooth, compliant 200Hz torque control—trained in 10 minutes!

Spending time around #NeurIPS2024 today! Ping to talk about (little) humanoids and (inverse) RL for sequential decision making!

Huge thanks to my collaborators: @MihaelaVDS @_chris_lu_ @RobertTLange @jfoerst.bsky.social @tennisonliu @QianZhaozhi @AlexJChan @claudio @weatheralljim75 🙌

DM me to connect at NeurIPS!

LLM agent automated scientific discovery for equation discovery/symbolic regression and active feature acquisition—towards an LLM scientist.
“Data-Driven Discovery of Dynamical Systems in Pharmacology using Large Language Models”
Main conference, Thu 12 Dec 4:30pm-7:30pm, East Exhibit Hall A-C #3202

LLM multi-agent automated neural architecture search/discovery through code representations consisting of equations and neural network components.
“Automatically Learning Hybrid Digital Twins of Dynamical Systems”
Main conference, Wed 11 Dec 11:00am-2:00pm, East Exhibit Hall A-C #3500

LLM automated objective discovery to discover a better offline preference optimization algorithm for LLM post-training (RLHF) to preferences.
“Discovering Preference Optimization Algorithms with and for Large Language Models”
Main conference, Thu 12 Dec 11:00am-2:00pm, East Exhibit Hall A-C #3304

Excited to be at #NeurIPS this week!
🚨I've wrapped up my internship at @GoogleDeepmind and am on the job market for RS positions/internships. I'm interested in LLM code generation/discovery, RL with LLMs (RLHF), RL/Control, LLM agents & reasoning. Will be presenting 3 papers at #NeurIPS2024 🧵: