Nvidia Corporation

University of California

Animesh Garg

Dieter Fox

Fabio Tozeto Ramos

Ken Goldberg

Sachin Patil

Andrew J Tao

Jonathan Tremblay

Bryan Catanzaro

Michael Garland

Miles Macklin

Stephen McKinley

Yu-Wei Chao

Kevin Shih

Christopher Jason Paxton

Balakumar Sundaralingam

Robert Pottorff

Jean Pouliot

Adithyavairavan Murali

Nathan Donald Ratliff

Byron Boots

Aysegul Dundar

Susan M L Lim

De-An Huang

Vinu Joseph

J Adam M Cunha

Pieter Abbeel

Timmy Siauw

Krishna Murthy Jatavallabhula

I-Chow Hsu

Yun-Chun Chen

Saurav Muralidharan

Ajay Uday Mandlekar

Michelle Lee

Carlos Florensa

Apparatuses, systems, and techniques to update a machine learning model associated with an object. In at least one embodiment, the machine learning model is updated based at least in part on, for example, one or more distributions associated with the machine learning model.

Predicting object models

One embodiment of a method includes determining, by a Bayesian optimizer, a first sparsity ratio associated with a limit on an accuracy loss caused by compressing the machine learning model. The method further includes selecting, by the Bayesian optimizer, a second sparsity ratio that optimizes a predefined objective function for the machine learning model within a search space bounded by the first sparsity ratio. The method further includes generating a compressed version of the machine learning model having the second sparsity ratio.

Bayesian optimization of sparsity ratios in model compression

One embodiment of a method for controlling a robot includes receiving sensor data associated with an environment that includes an object; applying a machine learning model to a portion of the sensor data associated with the object and one or more trajectories of motion of the robot to determine one or more path lengths of the one or more trajectories; generating a new trajectory of motion of the robot based on the one or more trajectories and the one or more path lengths; and causing the robot to perform one or more movements based on the new trajectory.

Techniques for robot control using neural implicit value functions

A robot is controlled using a combination of model-based and model-free control methods. In some examples, the model-based method uses a physical model of the environment around the robot to guide the robot. The physical model is oriented using a perception system such as a camera. Characteristics of the perception system may be are used to determine an uncertainty for the model. Based at least in part on this uncertainty, the system transitions from the model-based method to a model-free method where, in some embodiments, information provided directly from the perception system is used to direct the robot without reliance on the physical model.

Guided uncertainty-aware policy optimization: combining model-free and model-based strategies for sample-efficient learning

A framework for offline learning from a set of diverse and suboptimal demonstrations operates by selectively imitating local sequences from the dataset. At least one embodiment recovers performant policies from large manipulation datasets by decomposing the problem into a goal-conditioned imitation and a high-level goal selection mechanism.

Controlling position of robot by determining goal proposals by using neural networks

Apparatuses, systems, and techniques to enhance video are disclosed. In at least one embodiment, one or more neural networks are used to create, from a first video, a second video having one or more additional video frames.

Video prediction using one or more neural networks

Apparatuses, systems, and techniques to identify a goal of a demonstration. In at least one embodiment, video data of a demonstration is analyzed to identify a goal. Object trajectories identified in the video data are analyzed with respect to a task predicate satisfied by a respective object trajectory, and with respect to motion predicate. Analysis of the trajectory with respect to the motion predicate is used to assess intentionality of a trajectory with respect to the goal.

Imitation learning system

According to some embodiments of the invention, a surgical robot includes a robot arm having an end effector, the end effector comprising a needle assembly. The surgical robot further includes a robot control system operatively connected to the robot arm, and an end effector control system operatively connected to the end effector. The robot control system provides control signals for operation of the robot arm to move the end effector to selected positions relative to a subject. The end effector control system is configured to provide signals for operation of the end effector to at least one of inject material through the needle assembly to a selected location within the subject's body or extract material through the needle assembly from the selected location within the subject's body.

Precision injector/extractor for robot-assisted minimally-invasive surgery

The present invention offers an alternative for cancer treatment where radiation, thermotherapy, or another therapeutic modality must be delivered to an internal cavity of a subject, for example to treat mouth, anal, cervical, and vaginal cancers. The invention is a new approach that builds on recent results in 3D printing and steerable needle motion planning to create customized implants containing customized curvature-constrained internal channels that fit securely, minimize air gaps, and precisely guide treatment sources through internal printed channels to accurately reach tumors and minimize damage to healthy tissue.

Patient-specific temporary implants for accurately guiding local means of tumor control along patient-specific internal channels to treat cancer

Growing community of inventors

Berkeley, CA, United States of America

Animesh Garg