公开(公告)号：US20230044078A1

公开(公告)日：2023-02-09

申请号：US17816197

申请日：2022-07-29

Applicant: Google LLC

Inventor： Abhishek Kumar ,  Ehsan Amid

IPC: G06N20/00

Abstract: A method includes receiving training data for a machine learning model, the training data comprising a plurality of training examples and a corresponding plurality of labels. The method further includes dividing the training data into a plurality of training batches. For each training batch of the plurality of training batches, the method additionally includes learning a weight for each training example in the training batch that minimizes a sum of weighted losses for the training batch subject to a divergence constraint, where the divergence constraint limits a divergence of the learned weights for the training batch from a reference distribution, where the divergence is determined according to a chosen divergence measure. The method also includes training the machine learning model with each training batch of the plurality of training batches using the learned weight for each training example in the training batch. The method additionally includes providing the trained machine learning model.

公开(公告)号：US20230214663A1

公开(公告)日：2023-07-06

申请号：US17926282

申请日：2020-05-18

Applicant: Google LLC

Inventor： Abhishek Kumar ,  Esther Robb ,  Wen-Sheng Chu

IPC: G06N3/094 ,  G06N3/045

CPC classification number: G06N3/094 ,  G06N3/045

Abstract: The present disclosure provides improved methods for learning a generative model with limited training data, by leveraging a pre-trained GAN model from a related domain and adapting it to the new domain given a set of target examples from the new or target domain.

公开(公告)号：US20220398506A1

公开(公告)日：2022-12-15

申请号：US17837398

申请日：2022-06-10

Applicant: Google LLC

Inventor： Abhishek Kumar ,  Harikrishna Narasimhan ,  Andrew Steere Cotter

IPC: G06N20/20

Abstract: A computer-implemented method for optimizing machine-learned models by non-decomposable objectives with improved performance includes obtaining data indicative of a plurality of machine-learned model parameters and at least one threshold comprising a machine-learned model; initializing an initial plurality of machine-learned model parameters and an initial at least one threshold such that the initial plurality of machine-learned model parameters and the initial at least one threshold satisfy a constraint function; determining a gradient of an objective function with respect to the plurality of machine-learned model parameters at a current optimization step based at least in part on an implicit function of the at least one threshold as a function of the plurality of machine-learned model parameters; and updating the plurality of machine-learned model parameters and the at least one threshold based at least in part on the gradient.

公开(公告)号：US20250037353A1

公开(公告)日：2025-01-30

申请号：US18714862

申请日：2022-01-13

Applicant: Google LLC

Inventor： Wen-Sheng Chu ,  Dmitry Lagun ,  Ioannis Daras ,  Abhishek Kumar

IPC: G06T15/20 ,  G06N3/08

Abstract: Systems and methods for training a generative neural radiance field model can include geometric regularization. Geometric regularization can involve the utilization of reference geometry data and/or an output of a surface prediction model. The geometry regularization can train the generative neural radiance field model to mitigate artifact generation by limiting a distribution considered for color value prediction and density value prediction to a range associated with a realistic geometry range.

公开(公告)号：US12008821B2

公开(公告)日：2024-06-11

申请号：US17314738

申请日：2021-05-07

Applicant: Google LLC

Inventor： Wen-Sheng Chu ,  Abhishek Kumar ,  Min Jin Chong

IPC: G06K9/00 ,  G06F18/214 ,  G06F18/23 ,  G06N3/08 ,  G06N20/00 ,  G06T7/73 ,  G06V20/64 ,  G06V40/10 ,  G06V40/16

CPC classification number: G06V20/64 ,  G06F18/214 ,  G06F18/23 ,  G06N3/08 ,  G06N20/00 ,  G06T7/73 ,  G06V40/103 ,  G06V40/171

Abstract: Systems and methods of the present disclosure are directed to a computer-implemented method. The method can include obtaining a first image depicting a first object and a second image depicting a second object, wherein the first object comprises a first feature set and the second object comprises a second feature set. The method can include processing the first image with a machine-learned image transformation model comprising a plurality of model channels to obtain a first channel mapping indicative of a mapping between the plurality of model channels and the first feature set. The method can include processing the second image with the model to obtain a second channel mapping indicative of a mapping between the plurality of model channels and the second feature set. The method can include generating an interpolation vector for a selected feature.

公开(公告)号：US20220374625A1

公开(公告)日：2022-11-24

申请号：US17314738

申请日：2021-05-07

Applicant: Google LLC

Inventor： Wen-Sheng Chu ,  Abhishek Kumar ,  Min Jin Chong

IPC: G06K9/00 ,  G06N20/00 ,  G06K9/62 ,  G06T7/73 ,  G06N3/08

Abstract: Systems and methods of the present disclosure are directed to a computer-implemented method. The method can include obtaining a first image depicting a first object and a second image depicting a second object, wherein the first object comprises a first feature set and the second object comprises a second feature set. The method can include processing the first image with a machine-learned image transformation model comprising a plurality of model channels to obtain a first channel mapping indicative of a mapping between the plurality of model channels and the first feature set. The method can include processing the second image with the model to obtain a second channel mapping indicative of a mapping between the plurality of model channels and the second feature set. The method can include generating an interpolation vector for a selected feature.