公开(公告)号：US11423283B1

公开(公告)日：2022-08-23

申请号：US15933114

申请日：2018-03-22

Applicant: Amazon Technologies, Inc.

Inventor： Hagay Lupesko ,  Dominic Rajeev Divakaruni ,  Jonathan Esterhazy ,  Sandeep Krishnamurthy ,  Vikram Madan ,  Roshani Nagmote ,  Naveen Mysore Nagendra Swamy ,  Yao Wang

IPC: G06N3/04 ,  G06N3/08 ,  G06N3/10

Abstract: Techniques for model adaptation are described. For example, a method of receiving a call to provide either a model variant or a model variant profile of a deep learning model, the call including desired performance of the deep learning model, a deep learning model identifier, and current edge device characteristics; comparing the received current edge device characteristics to available model variants and profiles based on the desired performance of the deep learning model to generate or select a model variant or profile, the available model variants and profiles determined by the model identifier; and sending the generated or selected model variant or profile to the edge device to use in inference is detailed.

公开(公告)号：US10977518B1

公开(公告)日：2021-04-13

申请号：US16355617

申请日：2019-03-15

Applicant: Amazon Technologies, Inc.

Inventor： Rahul Sharma ,  Vikram Madan ,  James Robert Blair ,  Charles Bell

IPC: G06K9/62 ,  G06N5/04 ,  G06N20/00

Abstract: Techniques for generating and utilizing machine learning based adaptive instructions for annotation are described. An annotation service can use models to identify edge case data elements predicted to elicit differing annotations from annotators, “bad” data elements predicted to be difficult to annotate, and/or “good” data elements predicted to elicit matching or otherwise high-quality annotations from annotators. These sets of data elements can be automatically incorporated into annotation job instructions provided to annotators, resulting in improved overall annotation results via having efficiently and effectively “trained” the annotators how to perform the annotation task.

公开(公告)号：US11481906B1

公开(公告)日：2022-10-25

申请号：US16370723

申请日：2019-03-29

Applicant: Amazon Technologies, Inc.

Inventor： Hareesh Lakshmi Narayanan ,  Rahul Sharma ,  Arvind Jayasundar ,  Vikram Madan

IPC: G06T7/187 ,  G06K9/62 ,  G06N20/00 ,  G06V30/414

Abstract: Techniques for active learning-based data labeling are described. An active learning-based data labeling service enables a user to build and manage large, high accuracy datasets for use in various machine learning systems. Machine learning may be used to automate annotation and management of the datasets, increasing efficiency of labeling tasks and reducing the time required to perform labeling. Embodiments utilize active learning techniques to reduce the amount of a dataset that requires manual labeling. As subsets of the dataset are labeled, this label data is used to train a model which can then identify additional objects in the dataset without manual intervention. The process may continue iteratively until the model converges. This enables a dataset to be labeled without requiring each item in the data set to be individually and manually labeled by human labelers.

公开(公告)号：US10949252B1

公开(公告)日：2021-03-16

申请号：US15895747

申请日：2018-02-13

Applicant: Amazon Technologies, Inc.

Inventor： Sandeep Krishnamurthy ,  Jiajie Chen ,  Jonathan Esterhazy ,  Naveen Mysore Nagendra Swamy ,  Ruofei Yu ,  Yao Wang ,  Roshani Nagmote ,  Hagay Lupesko ,  Vikram Madan

IPC: G06F9/48 ,  G06N20/00

Abstract: Techniques for benchmarking a machine learning model/algorithm are described. For example, in some instances a method includes generating an execution plan for benchmarking of at least one task corresponding to a machine learning model based on an identified machine learning model, identified training data, and at least one objective for the benchmarking job; receiving execution statistics about the execution of the task as a part of the benchmarking job according to the execution plan; and updating the execution plan based at least in part on the received execution statistics of the task.