公开(公告)号：US10963754B1

公开(公告)日：2021-03-30

申请号：US16144927

申请日：2018-09-27

Applicant: Amazon Technologies, Inc.

Inventor： Avinash Aghoram Ravichandran ,  Paulo Ricardo dos Santos Mendonca ,  Rahul Bhotika ,  Stefano Soatto

IPC: G06K9/00 ,  G06K9/62 ,  G06N3/08 ,  G06N3/04

Abstract: Techniques for training an embedding using a limited training set are described. In some examples, the embedding is trained by generating a plurality of vectors from a random sample of the limited set of training data classes using a layer of the particular machine learning classification model, randomly selecting samples from the plurality of vectors into a set of samples, computing at least one distance for each sampled class from a center parameter for the class using the set of samples, generating a discrete probability distribution over the classes for a query point based on distances to a center parameter for each of the classes in the embedding space, calculating a loss value for the modified prototypical network, the calculation of the loss value being for a fixed geometry of the embedding space and including a measure of the difference between distributions, and back propagating.

公开(公告)号：US12182673B1

公开(公告)日：2024-12-31

申请号：US17315110

申请日：2021-05-07

Applicant: Amazon Technologies, Inc.

Inventor： Shuo Yang ,  Hao Zhou ,  Yuanjun Xiong ,  Wei Xia ,  Stefano Soatto

IPC: G06N3/04 ,  G06N20/00

Abstract: Machine learning models may be generated that are compatible with another machine learning model and satisfy a resource constraint. Techniques that ensure weight compatibility and architectural compatibility between a machine learning model being created to be compatible with another machine learning model are applied. The resource constraint is enforced so that the generated machine learning model also fits within the resource constraint.

公开(公告)号：US11860977B1

公开(公告)日：2024-01-02

申请号：US17307701

申请日：2021-05-04

Applicant: Amazon Technologies, Inc.

Inventor： Yifan Xing ,  Tianjun Xiao ,  Tong He ,  Yongxin Wang ,  Yuanjun Xiong ,  Wei Xia ,  David Paul Wipf ,  Zheng Zhang ,  Stefano Soatto

IPC: G06F18/2323 ,  G06N20/00 ,  G06F18/2415 ,  G06F18/23213 ,  G06F18/2413

CPC classification number: G06F18/2323 ,  G06F18/23213 ,  G06F18/2415 ,  G06F18/24147 ,  G06N20/00

Abstract: Techniques for performing visual clustering with a hierarchical graph neural network framework including a joint linkage prediction and density estimation graph model are described. Embodiments herein recurrently run the joint linkage prediction and density estimation graph model to generate intermediate clusters in multiple iterations (e.g., until convergence) to obtain a final clustering result. In certain embodiments, for each iteration, the input graph contains nodes that are merged from nodes assigned to intermediate clusters from the previous iteration. By using a small and fixed bandwidth k in each iteration, embodiments herein alleviate the sensitivity to the k selection for different clustering applications. Certain embodiments herein remove the tuning of a different k (e.g., k-bandwidth) for k-nearest neighbor graph construction over different clustering applications.

公开(公告)号：US11257006B1

公开(公告)日：2022-02-22

申请号：US16196662

申请日：2018-11-20

Applicant: Amazon Technologies, Inc.

Inventor： Oron Anschel ,  Amit Adam ,  Shahar Tsiper ,  Hadar Averbuch Elor ,  Shai Mazor ,  Rahul Bhotika ,  Stefano Soatto

IPC: G06N20/00 ,  G06K9/00 ,  G06F40/169

Abstract: Techniques for auto-generation of annotated real-world training data are described. An electronic document is analyzed to determine text represented in the document and corresponding locations of the text. A representation of the electronic document is modified to include markers and printed. The printed document is photographed in real-world environments, and the markers within the digital photographs are analyzed to allow for the depiction of the document within the photographs to be rectified. The text and location data are used to annotate the rectified images.

公开(公告)号：US12217137B1

公开(公告)日：2025-02-04

申请号：US17039447

申请日：2020-09-30

Applicant: Amazon Technologies, Inc.

Inventor： Rasool Fakoor ,  Alexander Johannes Smola ,  Stefano Soatto ,  Pratik Anil Chaudhari

IPC: G06N20/00

Abstract: Techniques for Meta-Q-Learning (MQL) are described. A method of MQL may include receiving a request from an agent to perform adaptation based at least on task data associated with a new task collected by the agent, identifying a subset of meta-training data corresponding to the task data in a replay buffer, and adapting a policy using the subset of meta-training data and the task data to generate an adapted policy, wherein the adapted policy is used identify a next action for the agent to perform.

公开(公告)号：US11429813B1

公开(公告)日：2022-08-30

申请号：US16697662

申请日：2019-11-27

Applicant: Amazon Technologies, Inc.

Inventor： Avinash Aghoram Ravichandran ,  Rahul Bhotika ,  Stefano Soatto ,  Pietro Perona ,  Hao Yang

IPC: G06K9/62 ,  G06N20/00

Abstract: This disclosure describes automatically selecting and training one or more models for image recognition based upon training and testing (validation) data provided by a user. A service provider network includes a recognition service that may use models to process images and videos to recognize objects in the images and videos, features on the objects in the images and videos, and/or locate objects in the images and videos. The service provider network also includes a model selection and training service that may select one or more modeling techniques based on the objectives of the user and/or the amount of data provided by the user. Based on the selected modeling technique, the model selection and training service selects and trains one or more models for use by the recognition service to process images and videos using the training data. The trained model may be tested and validated using the testing data.

公开(公告)号：US11216697B1

公开(公告)日：2022-01-04

申请号：US16815787

申请日：2020-03-11

Applicant: Amazon Technologies, Inc.

Inventor： Yantao Shen ,  Yuanjun Xiong ,  Siqi Deng ,  Wei Xia ,  Shuo Yang ,  Yifan Xing ,  Wei Li ,  Stefano Soatto

IPC: G06K9/62 ,  G06K9/00 ,  G06N20/00 ,  G06F16/538

Abstract: Techniques for building a backward compatible and backfill-free image search system are described. According to some embodiments, a backwards compatible training system trains a new embedding model to be backward compatible with the face embeddings (e.g., floating-point vectors) generated by a previous embedding model. In one embodiment, backwards compatible training uses a classifier of the previous embedding model as a form of constraint in the training of the new embedding model.

公开(公告)号：US10970530B1

公开(公告)日：2021-04-06

申请号：US16189633

申请日：2018-11-13

Applicant: Amazon Technologies, Inc.

Inventor： Amit Adam ,  Oron Anschel ,  Or Perel ,  Gal Sabina Star ,  Omri Ben-Eliezer ,  Hadar Averbuch Elor ,  Shai Mazor ,  Wendy Tse ,  Andrea Olgiati ,  Rahul Bhotika ,  Stefano Soatto

IPC: G06K9/62 ,  G06F40/137 ,  G06F40/169 ,  G06F40/174 ,  G06K9/00 ,  G06N20/00

Abstract: Techniques for grammar-based automated generation of annotated synthetic form training data for machine learning are described. A training data generation engine utilizes a defined grammar to construct a layout for a form, select key-value units to place within the layout, and select attribute variants for the key-value units. The form is rendered and stored at a storage location, where it can be provided along with other similarly-generated forms to be used as training data for a machine learning model.

公开(公告)号：US10762644B1

公开(公告)日：2020-09-01

申请号：US16218973

申请日：2018-12-13

Applicant: Amazon Technologies, Inc.

Inventor： Vijay Mahadevan ,  Stefano Soatto

IPC: G06K9/00 ,  G06T7/246 ,  G06N3/04 ,  G06N3/08 ,  G06F17/18 ,  G06T11/20

Abstract: Techniques for multiple object tracking in video are described in which the outputs of neural networks are combined within a Bayesian framework. A motion model is applied to a probability distribution representing the estimated current state of a target object being tracked to predict the state of the target object in the next frame. A state of an object can include one or more features, such as the location of the object in the frame, a velocity and/or acceleration of the object across frames, a classification of the object, etc. The prediction of the state of the target object in the next frame is adjusted by a score based on the combined outputs of neural networks that process the next frame.