Apple Inc.

Carnegie Mellon University

Robert Boschgmbh

Ruslan Salakhutdinov

Nitish Srivastava

Yichuan Tang

Jian Zhang

Hanlin Goh

Emilio Parisotto

Devendra Singh Chaplot

Filipe J Cabrita Condessa

Jeremy Zieg Kolter

João D Semedo

Mario J Srouji

Bahare Azari

Yutong He

Rui Hu

Marcus Pereira

A method discloses receiving, at a cross-attention layer of a model, first text data describing a first object and second text data describing a first scene, wherein the first text data includes a description of a location of the first object, utilizing the model with cross-attention layers, concatenating the first text data and the second text data to generate a prompt; generating, a broadcasted location mask constructed from at least the location; generating, a broadcasted all-one matrix associated with the second text data described the first scene; computing a key matrix and a value matrix utilizing separate linear projections of the prompt; computing a query matrix utilizing linear projections; generating a broadcasted location matrix in response to concatenating the broadcasted location mask and the broadcasted all-one matrix; generating a cross-attention map utilizing the query matrix, the key matrix, and the broadcasted location matrix; and outputting a final image.

System and method for localized text to image generation for free via cross attention

A method includes defining a geometric capsule that is interpretable by a capsule neural network, wherein the geometric capsule includes a feature representation and a pose. The method also includes determining multiple viewpoints relative to the geometric capsule and determining a first appearance representation of the geometric capsule for each of the multiple viewpoints. The method also includes determining a transform for each of the multiple viewpoints that moves each of the multiple viewpoints to a respective transformed viewpoint and determining second appearance representations that each correspond to one of the transformed viewpoints. The method also includes combining the second appearance representations to define an agreed appearance representation. The method also includes updating the feature representation for the geometric capsule based on the agreed appearance representation.

Encoding three-dimensional data for processing by capsule neural networks

A method includes obtaining a two-dimensional image, obtaining a two-dimensional image annotation that indicates presence of an object in the two-dimensional image, obtaining three-dimensional sensor information, generating a top-down representation of the three-dimensional sensor information, and obtaining a top-down annotation that indicates presence of the object in the top-down representation. The method also includes determining a bottom surface of a three-dimensional cuboid based on map information, determining a position, a length, a width, and a yaw rotation of the three-dimensional cuboid based on the top-down annotation, and determining a height of the three-dimensional cuboid based on a two-dimensional image annotation, and the position, the length, the width, and the yaw rotation of the three-dimensional cuboid.

Three-dimensional object estimation using two-dimensional annotations

At a centralized model trainer, one or more neural network based models are trained using an input data set. At least a first set of parameters of a model is transmitted to a model deployment destination. Using a second input data set, one or more adaptive parameters for the model are determined at the model deployment destination. Using the adaptive parameters, one or more inferences are generated at the model deployment destination.

Adaptive training of neural network models at model deployment destinations

A method includes determining motion imitation information for causing a system to imitate a physical task using a first machine learning model that is trained using motion information that represents a performance of the physical task, determining a predicted correction based on the motion information and a current state from the system using a second machine learning model that is trained using the motion information, determining an action to be performed by the system based on the motion imitation information and the predicted correction; and controlling motion of the system in accordance with the action.

Learning skills from video demonstrations

A computer-implemented system and method for storing data associated with an agent in a multi-dimensional environment via a memory architecture. The memory architecture is structured so that each unique position in the environment corresponds to a unique position within the memory architecture, thereby allowing the memory architecture to store features located at a particular position in the environment in a memory location specific to that location. As the agent traverses the environment, the agent compares the features at the agent's particular position to a summary of the features stored throughout the memory architecture and writes the features that correspond to the summary to the coordinates in the memory architecture that correspond to the agent's position. The system and method allows agents to learn, using a reinforcement signal, how to behave when acting in an environment that requires storing information over long time steps.

Neural map

Sensor data captured by one or more sensors may be received at an analysis system. A neural network may be used to detect an object in the sensor data. A plurality of polygons surrounding the object may be generated in one or more subsets of the sensor data. A prediction of a future position of the object may be generated based at least in part on the polygons. One or more commands may be provided to a control system based on the prediction of the future position.

Analysis of objects of interest in sensor data using deep neural networks

A method includes obtaining a two-dimensional image, obtaining a two-dimensional image annotation that indicates presence of an object in the two-dimensional image, determining a location proposal based on the two-dimensional image annotation, determining a classification for the object, determining an estimated size for the object based on the classification for the object, and defining a three-dimensional cuboid for the object based on the location proposal and the estimated size.

In accordance with some embodiments, a method is performed at a device with one or more processors and non-transitory memory. The method includes obtaining location vector data characterizing an object. The method includes determining a neural pose graph associated with a respective time-period based on an initial local pose estimation as a function of respective location vector data. The method includes determining a meta pose estimation associated with the respective time-period by aggregating the neural pose graph associated with the respective time-period and one or more other neural pose graphs associated with one or more temporally adjacent time-periods. The method includes synthesizing a corrected pose estimation by correcting the meta pose estimation associated with the respective time-period based on a function of the meta pose estimation associated with the respective time-period and meta pose estimations associated with one or more temporally adjacent time-periods in order to correct accumulated errors in the initial local pose estimation.

Method and device for improved localization and mapping

A system employs an inspection neural network (INN) to inspect data generated during an inference process of a primary neural network (PNN) to generate an indication of reliability for an output generated by the PNN. The system includes a sensor configured to capture sensor data. Sensor data captured by the sensor is provided to a data analyzer to generate an output using the PNN. An analyzer inspector is configured to capture inspection data associated with the generation of the output by the data analyzer, and use the INN to generate an indication of reliability for the PNN's output based on the inspection data. The INN is trained using a set of training data that is distinct from the training data used to train the PNN.

Growing community of inventors

Pittsburgh, PA, United States of America

Ruslan Salakhutdinov