Mrinal Kalakrishnan

TL;DR: QT-Opt as mentioned in this paper is a scalable self-supervised vision-based reinforcement learning framework that can leverage over 580k real-world grasp attempts to train a deep neural network Q-function with over 1.2M parameters.

TL;DR: It is experimentally show that the stochastic nature of STOMP allows it to overcome local minima that gradient-based methods like CHOMP can get stuck in.

TL;DR: In this paper, the authors study how randomized simulated environments and domain adaptation methods can be extended to train a grasping system to grasp novel objects from raw monocular RGB images, and they extensively evaluate their approaches with a total of more than 25,000 physical test grasps, including a novel extension of pixel-level domain adaptation that they termed the GraspGAN.

TL;DR: This paper presents Randomized-to-Canonical Adaptation Networks (RCANs), a novel approach to crossing the visual reality gap that uses no real-world data and learns to translate randomized rendered images into their equivalent non-randomized, canonical versions.

TL;DR: A floating-base inverse dynamics controller that allows for robust, compliant locomotion over unperceived obstacles and the generalization ability of this controller is demonstrated by presenting results from testing performed by an independent external test team on terrain that has never been shown to us.