-
公开(公告)号：US20190354857A1
公开(公告)日：2019-11-21
申请号：US16413730
申请日：2019-05-16
Applicant: Raytheon Company
Inventor： Philip A. Sallee , James Mullen , Franklin Tanner
Abstract: Subject matter regards improving machine learning techniques using informed pseudolabels. A method can include receiving previously assigned labels indicating an expected classification for data, the labels having a specified uncertainty, generating respective pseudolabels for the data based on the previously assigned labels, the data, a class vector determined by an ML model, and a noise model indicating, based on the specified uncertainty, a likelihood of the previously assigned label given the class, and substituting the pseudolabels for the previously assigned labels in a next epoch of training the ML model.
-
公开(公告)号：US20210223777A1
公开(公告)日：2021-07-22
申请号：US16745885
申请日：2020-01-17
Applicant: RAYTHEON COMPANY
Inventor： Rupal Nigam , James Mullen
Abstract: Techniques for multi-factor pathfinding are disclosed. A system receives multiple data sets that represent multiple physical properties associated with multiple three-dimensional coordinates of a physical environment. The system generates multiple weighted data layers corresponding to the physical properties, at least by applying weights to the data sets. The system generates a multi-factor costmap for the three-dimensional environment, based at least on the weighted data layers. The system determines a path through the three-dimensional environment, based at least on the multi-factor costmap.
-
公开(公告)号：US11726482B2
公开(公告)日：2023-08-15
申请号：US16745885
申请日：2020-01-17
Applicant: RAYTHEON COMPANY
Inventor： Rupal Nigam , James Mullen
CPC classification number: G05D1/0212 , G01C21/3676 , G05D1/0274 , G06F16/29
Abstract: Techniques for multi-factor pathfinding are disclosed. A system receives multiple data sets that represent multiple physical properties associated with multiple three-dimensional coordinates of a physical environment. The system generates multiple weighted data layers corresponding to the physical properties, at least by applying weights to the data sets. The system generates a multi-factor costmap for the three-dimensional environment, based at least on the weighted data layers. The system determines a path through the three-dimensional environment, based at least on the multi-factor costmap.
-
Patent Agency Ranking
公开(公告)号：US11068747B2
公开(公告)日：2021-07-20
申请号：US16586480
申请日：2019-09-27
Applicant: Raytheon Company
Inventor： Jonathan Goldstein , Philip A. Sallee , James Mullen , Franklin Tanner
Abstract: A neural network apparatus includes processing circuitry and memory. The memory stores a plurality of images of a target. The processing circuitry is configured to: access, from the memory, a first image and an identification of a centroid pixel of the target within the first image; generate, based on a geometry of the target and the centroid pixel, a confidence map indicating, for each pixel in the first image, a confidence value that the pixel includes the target; train, using the plurality of images of the target, including the first image and the confidence map, an artificial neural network to identify the target in visual data; and provide an output representing the trained artificial neural network.
-
公开(公告)号：US20210097345A1
公开(公告)日：2021-04-01
申请号：US16586480
申请日：2019-09-27
Applicant: Raytheon Company
Inventor： Jonathan Goldstein , Philip A. Sallee , James Mullen , Franklin Tanner
Abstract: A neural network apparatus includes processing circuitry and memory. The memory stores a plurality of images of a target. The processing circuitry is configured to: access, from the memory, a first image and an identification of a centroid pixel of the target within the first image; generate, based on a geometry of the target and the centroid pixel, a confidence map indicating, for each pixel in the first image, a confidence value that the pixel includes the target; train, using the plurality of images of the target, including the first image and the confidence map, an artificial neural network to identify the target in visual data; and provide an output representing the trained artificial neural network.
-
公开(公告)号：US12141238B2
公开(公告)日：2024-11-12
申请号：US17081612
申请日：2020-10-27
Applicant: Raytheon Company
Inventor： Philip A. Sallee , James Mullen
IPC: G06N3/063 , G06F18/22 , G06F18/2321 , G06F18/2415 , G06N3/04 , G06N3/08
Abstract: Discussed herein are devices, systems, and methods for classification using a clustering autoencoder. A method can include obtaining content to be classified by the DNN classifier, and operating the DNN classifier to determine a classification of the received content, the DNN classifier including a clustering classification layer that clusters based on a latent feature vector representation of the content, the classification corresponding to one or more clusters that are closest to the latent feature vector providing the classification and a corresponding confidence.
-
公开(公告)号：US11669724B2
公开(公告)日：2023-06-06
申请号：US16413730
申请日：2019-05-16
Applicant: Raytheon Company
Inventor： Philip A. Sallee , James Mullen , Franklin Tanner
Abstract: Subject matter regards improving machine learning techniques using informed pseudolabels. A method can include receiving previously assigned labels indicating an expected classification for data, the labels having a specified uncertainty, generating respective pseudolabels for the data based on the previously assigned labels, the data, a class vector determined by an ML model, and a noise model indicating, based on the specified uncertainty, a likelihood of the previously assigned label given the class, and substituting the pseudolabels for the previously assigned labels in a next epoch of training the ML model.
-
公开(公告)号：US20220129712A1
公开(公告)日：2022-04-28
申请号：US17081612
申请日：2020-10-27
Applicant: Raytheon Company
Inventor： Philip A. Sallee , James Mullen
Abstract: Discussed herein are devices, systems, and methods for classification using a clustering autoencoder. A method can include obtaining content to be classified by the DNN classifier, and operating the DNN classifier to determine a classification of the received content, the DNN classifier including a clustering classification layer that clusters based on a latent feature vector representation of the content, the classification corresponding to one or more clusters that are closest to the latent feature vector providing the classification and a corresponding confidence.
-
-
-
-
-
-
-
Search Results
8
records
(took 0.013 seconds)
