Georgia Chalvatzaki

SteinSQP achieves robust and diverse motion planning by ensuring all sampled trajectories are feasible while significantly speeding up convergence compared to traditional methods.

Synthetic data generation via RL not only scales but also enhances generalization in bimanual dexterous manipulation by leveraging language-conditioned task annotations.

Automating the RL workflow with HARBOR cuts engineering costs and enhances policy transferability to real-world robots.

Multi-resolution tactile sensing boosts robotic manipulation success rates to 80%, far surpassing traditional vision-only approaches.

Standard LLMs can now perform complex bimanual robot manipulation tasks with impressive success rates, all without any task-specific training.

Surprisingly, even a mediocre whole-body controller can be leveraged with offline RL to achieve impressive mobile manipulation, rivaling hand-tuned controllers and generalizing to the real world without any real-world training.

Forget expensive real-world data collection: a massive, diverse synthetic dataset enables surprisingly effective zero-shot transfer for robotic manipulation.

Robots can now achieve superior surface coverage with precise end-effector poses thanks to a new SE(3)-aware Stein Variational Gradient Descent method that outperforms existing trajectory optimization techniques.

By distilling a frozen diffusion model's geometric understanding into a fast, deterministic network, Robot-DIFT unlocks more precise robot control compared to standard vision encoders.

Forget synthetic benchmarks that don't translate: MolmoSpaces offers 230k diverse, simulator-agnostic environments with 130k annotated objects, showing a remarkable 0.96 sim-to-real correlation for robot policies.