Abstract: Systems and methods for parsing guideline data to generate training data for generative models. The method can include accessing guideline data and parsing the guideline data to generate a number of discrete criteria. The method can include accessing content item data and determining, for each discrete criteria of the number of discrete criteria, a label for the first content item of at least one of (i) satisfies or (ii) violates. The method can include transmitting instructions that are executable to provide the first content item along with a label-guideline pair for each of the number of discrete criteria. The method can include accessing data including a user altering a label-guideline pair for a first discrete guideline. The method can include updating a training dataset to include the label-guideline pair for the first discrete guideline. The method can include training an asset generation model using the updated training dataset.

Abstract: Mitigating the reality gap through optimization of one or more simulated hardware parameters for simulated hardware components of a simulated robot. Implementations generate and store real navigation data instances that are each based on a corresponding episode of locomotion of a real robot. A real navigation data instance can include a sequence of velocity control instances generated to control a real robot during a real episode of locomotion of the real robot, and one or more ground truth values, where each of the ground truth values is a measured value of a corresponding property of the real robot (e.g., pose). The velocity control instances can be applied to a simulated robot, and one or more losses can be generated based on comparing the ground truth value(s) to corresponding simulated value(s) generated from applying the velocity control instances to the simulated robot. The simulated hardware parameters and environmental parameters can be optimized based on the loss(es).

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for determining neural network architectures.

Abstract: Mitigating the reality gap through optimization of one or more simulated hardware parameters for simulated hardware components of a simulated robot. Implementations generate and store real navigation data instances that are each based on a corresponding episode of locomotion of a real robot. A real navigation data instance can include a sequence of velocity control instances generated to control a real robot during a real episode of locomotion of the real robot, and one or more ground truth values, where each of the ground truth values is a measured value of a corresponding property of the real robot (e.g., pose). The velocity control instances can be applied to a simulated robot, and one or more losses can be generated based on comparing the ground truth value(s) to corresponding simulated value(s) generated from applying the velocity control instances to the simulated robot. The simulated hardware parameters and environmental parameters can be optimized based on the loss(es).