Patents by Inventor Abhishek Sinha
Abhishek Sinha has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).
-
Patent number: 11972466Abstract: A search system provides search results with images of products based on associations of primary products and secondary products from product image sets. The search system analyzes a product image set containing multiple images to determine a primary product and secondary products. Information associating the primary and secondary products are stored in a search index. When the search system receives a query image containing a search product, the search index is queried using the search product to identify search result images based on associations of products in the search index, and the result images are provided as a response to the query image.Type: GrantFiled: May 20, 2019Date of Patent: April 30, 2024Assignee: ADOBE INCInventors: Jonas Dahl, Mausoom Sarkar, Hiresh Gupta, Balaji Krishnamurthy, Ayush Chopra, Abhishek Sinha
-
Publication number: 20240086138Abstract: In accordance with some embodiments, the render rate is varied across and/or up and down the display screen. This may be done based on where the user is looking in order to reduce power consumption and/or increase performance. Specifically the screen display is separated into regions, such as quadrants. Each of these regions is rendered at a rate determined by at least one of what the user is currently looking at, what the user has looked at in the past and/or what it is predicted that the user will look at next. Areas of less focus may be rendered at a lower rate, reducing power consumption in some embodiments.Type: ApplicationFiled: September 26, 2023Publication date: March 14, 2024Inventors: Eric J. Asperheim, Subramaniam Maiyuran, Kiran C. Veernapu, Sanjeev S. Jahagirdar, Balaji Vembu, Devan Burke, Philip R. Laws, Kamal Sinha, Abhishek R. Appu, Elmoustapha Ould-Ahmed-Vall, Peter L. Doyle, Joydeep Ray, Travis T. Schluessler, John H. Feit, Nikos Kaburlasos, Jacek Kwiatkowski, Altug Koker
-
Patent number: 11922535Abstract: Embodiments provide mechanisms to facilitate compute operations for deep neural networks. One embodiment comprises a graphics processing unit comprising one or more multiprocessors, at least one of the one or more multiprocessors including a register file to store a plurality of different types of operands and a plurality of processing cores. The plurality of processing cores includes a first set of processing cores of a first type and a second set of processing cores of a second type. The first set of processing cores are associated with a first memory channel and the second set of processing cores are associated with a second memory channel.Type: GrantFiled: February 13, 2023Date of Patent: March 5, 2024Assignee: Intel CorporationInventors: Prasoonkumar Surti, Narayan Srinivasa, Feng Chen, Joydeep Ray, Ben J. Ashbaugh, Nicolas C. Galoppo Von Borries, Eriko Nurvitadhi, Balaji Vembu, Tsung-Han Lin, Kamal Sinha, Rajkishore Barik, Sara S. Baghsorkhi, Justin E. Gottschlich, Altug Koker, Nadathur Rajagopalan Satish, Farshad Akhbari, Dukhwan Kim, Wenyin Fu, Travis T. Schluessler, Josh B. Mastronarde, Linda L. Hurd, John H. Feit, Jeffery S. Boles, Adam T. Lake, Karthik Vaidyanathan, Devan Burke, Subramaniam Maiyuran, Abhishek R. Appu
-
Patent number: 11829880Abstract: The present disclosure relates to systems, methods, and non-transitory computer readable media for generating trained neural network with increased robustness against adversarial attacks by utilizing a dynamic dropout routine and/or a cyclic learning rate routine. For example, the disclosed systems can determine a dynamic dropout probability distribution associated with neurons of a neural network. The disclosed systems can further drop neurons from a neural network based on the dynamic dropout probability distribution to help neurons learn distinguishable features. In addition, the disclosed systems can utilize a cyclic learning rate routine to force copy weights of a copy neural network away from weights of an original neural network without decreasing prediction accuracy to ensure that the decision boundaries learned are different.Type: GrantFiled: October 24, 2022Date of Patent: November 28, 2023Assignee: Adobe Inc.Inventors: Mayank Singh, Nupur Kumari, Dhruv Khattar, Balaji Krishnamurthy, Abhishek Sinha
-
Patent number: 11734337Abstract: The present disclosure relates to systems, methods, and non-transitory computer-readable media for generating tags for an object portrayed in a digital image based on predicted attributes of the object. For example, the disclosed systems can utilize interleaved neural network layers of alternating inception layers and dilated convolution layers to generate a localization feature vector. Based on the localization feature vector, the disclosed systems can generate attribute localization feature embeddings, for example, using some pooling layer such as a global average pooling layer. The disclosed systems can then apply the attribute localization feature embeddings to corresponding attribute group classifiers to generate tags based on predicted attributes. In particular, attribute group classifiers can predict attributes as associated with a query image (e.g., based on a scoring comparison with other potential attributes of an attribute group).Type: GrantFiled: June 14, 2022Date of Patent: August 22, 2023Assignee: Adobe Inc.Inventors: Ayush Chopra, Mausoom Sarkar, Jonas Dahl, Hiresh Gupta, Balaji Krishnamurthy, Abhishek Sinha
-
Patent number: 11734565Abstract: Embodiments disclosed herein describe systems, methods, and products that generate trained neural networks that are robust against adversarial attacks. During a training phase, an illustrative computer may iteratively optimize a loss function that may include a penalty for ill-conditioned weight matrices in addition to a penalty for classification errors. Therefore, after the training phase, the trained neural network may include one or more well-conditioned weight matrices. The one or more well-conditioned weight matrices may minimize the effect of perturbations within an adversarial input thereby increasing the accuracy of classification of the adversarial input. By contrast, conventional training approaches may merely reduce the classification errors using backpropagation, and, as a result, any perturbation in an input is prone to generate a large effect on the output.Type: GrantFiled: June 3, 2022Date of Patent: August 22, 2023Assignee: Adobe Inc.Inventors: Mayank Singh, Abhishek Sinha, Balaji Krishnamurthy
-
Publication number: 20230107574Abstract: The present disclosure relates to systems, methods, and non-transitory computer readable media for generating trained neural network with increased robustness against adversarial attacks by utilizing a dynamic dropout routine and/or a cyclic learning rate routine. For example, the disclosed systems can determine a dynamic dropout probability distribution associated with neurons of a neural network. The disclosed systems can further drop neurons from a neural network based on the dynamic dropout probability distribution to help neurons learn distinguishable features. In addition, the disclosed systems can utilize a cyclic learning rate routine to force copy weights of a copy neural network away from weights of an original neural network without decreasing prediction accuracy to ensure that the decision boundaries learned are different.Type: ApplicationFiled: October 24, 2022Publication date: April 6, 2023Inventors: Mayank Singh, Nupur Kumari, Dhruv Khattar, Balaji Krishnamurthy, Abhishek Sinha
-
Patent number: 11481617Abstract: The present disclosure relates to systems, methods, and non-transitory computer readable media for generating trained neural network with increased robustness against adversarial attacks by utilizing a dynamic dropout routine and/or a cyclic learning rate routine. For example, the disclosed systems can determine a dynamic dropout probability distribution associated with neurons of a neural network. The disclosed systems can further drop neurons from a neural network based on the dynamic dropout probability distribution to help neurons learn distinguishable features. In addition, the disclosed systems can utilize a cyclic learning rate routine to force copy weights of a copy neural network away from weights of an original neural network without decreasing prediction accuracy to ensure that the decision boundaries learned are different.Type: GrantFiled: January 22, 2019Date of Patent: October 25, 2022Assignee: Adobe Inc.Inventors: Mayank Singh, Nupur Kumari, Dhruv Khattar, Balaji Krishnamurthy, Abhishek Sinha
-
Patent number: 11468314Abstract: Embodiments disclosed herein describe systems, methods, and products that generate trained neural networks that are robust against adversarial attacks. During a training phase, an illustrative computer may iteratively optimize a loss function that may include a penalty for ill-conditioned weight matrices in addition to a penalty for classification errors. Therefore, after the training phase, the trained neural network may include one or more well-conditioned weight matrices. The one or more well-conditioned weight matrices may minimize the effect of perturbations within an adversarial input thereby increasing the accuracy of classification of the adversarial input. By contrast, conventional training approaches may merely reduce the classification errors using backpropagation, and, as a result, any perturbation in an input is prone to generate a large effect on the output.Type: GrantFiled: September 12, 2018Date of Patent: October 11, 2022Assignee: Adobe Inc.Inventors: Mayank Singh, Abhishek Sinha, Balaji Krishnamurthy
-
Publication number: 20220309093Abstract: The present disclosure relates to systems, methods, and non-transitory computer-readable media for generating tags for an object portrayed in a digital image based on predicted attributes of the object. For example, the disclosed systems can utilize interleaved neural network layers of alternating inception layers and dilated convolution layers to generate a localization feature vector. Based on the localization feature vector, the disclosed systems can generate attribute localization feature embeddings, for example, using some pooling layer such as a global average pooling layer. The disclosed systems can then apply the attribute localization feature embeddings to corresponding attribute group classifiers to generate tags based on predicted attributes. In particular, attribute group classifiers can predict attributes as associated with a query image (e.g., based on a scoring comparison with other potential attributes of an attribute group).Type: ApplicationFiled: June 14, 2022Publication date: September 29, 2022Inventors: Ayush Chopra, Mausoom Sarkar, Jonas Dahl, Hiresh Gupta, Balaji Krishnamurthy, Abhishek Sinha
-
Publication number: 20220292356Abstract: Embodiments disclosed herein describe systems, methods, and products that generate trained neural networks that are robust against adversarial attacks. During a training phase, an illustrative computer may iteratively optimize a loss function that may include a penalty for ill-conditioned weight matrices in addition to a penalty for classification errors. Therefore, after the training phase, the trained neural network may include one or more well-conditioned weight matrices. The one or more well-conditioned weight matrices may minimize the effect of perturbations within an adversarial input thereby increasing the accuracy of classification of the adversarial input. By contrast, conventional training approaches may merely reduce the classification errors using backpropagation, and, as a result, any perturbation in an input is prone to generate a large effect on the output.Type: ApplicationFiled: June 3, 2022Publication date: September 15, 2022Inventors: Mayank SINGH, Abhishek SINHA, Balaji KRISHNAMURTHY
-
Patent number: 11386144Abstract: The present disclosure relates to systems, methods, and non-transitory computer-readable media for generating tags for an object portrayed in a digital image based on predicted attributes of the object. For example, the disclosed systems can utilize interleaved neural network layers of alternating inception layers and dilated convolution layers to generate a localization feature vector. Based on the localization feature vector, the disclosed systems can generate attribute localization feature embeddings, for example, using some pooling layer such as a global average pooling layer. The disclosed systems can then apply the attribute localization feature embeddings to corresponding attribute group classifiers to generate tags based on predicted attributes. In particular, attribute group classifiers can predict attributes as associated with a query image (e.g., based on a scoring comparison with other potential attributes of an attribute group).Type: GrantFiled: September 9, 2019Date of Patent: July 12, 2022Assignee: Adobe Inc.Inventors: Ayush Chopra, Mausoom Sarkar, Jonas Dahl, Hiresh Gupta, Balaji Krishnamurthy, Abhishek Sinha
-
Patent number: 11308353Abstract: The present disclosure relates to systems, methods, and non-transitory computer readable media for training a classification neural network to classify digital images in few-shot tasks based on self-supervision and manifold mixup. For example, the disclosed systems can train a feature extractor as part of a base neural network utilizing self-supervision and manifold mixup. Indeed, the disclosed systems can apply manifold mixup regularization over a feature manifold learned via self-supervised training such as rotation training or exemplar training. Based on training the feature extractor, the disclosed systems can also train a classifier to classify digital images into novel classes not present within the base classes used to train the feature extractor.Type: GrantFiled: October 23, 2019Date of Patent: April 19, 2022Assignee: Adobe Inc.Inventors: Mayank Singh, Puneet Mangla, Nupur Kumari, Balaji Krishnamurthy, Abhishek Sinha
-
Patent number: 11109380Abstract: Methods, systems, and devices for wireless communications are described. One method may include receiving, by a user equipment (UE) in a downlink shared data channel, a beam switch command instructing the UE to switch from a first uplink control beam to a second uplink control beam. The UE may select an uplink control beam to transmit acknowledgment feedback based on whether decoding the beam switch command is successful, and transmit, in an uplink control channel via the selected uplink control beam, the acknowledgment feedback indicating whether decoding the beam switch command was successful.Type: GrantFiled: March 4, 2019Date of Patent: August 31, 2021Assignee: QUALCOMM IncorporatedInventors: Makesh Pravin John Wilson, Sony Akkarakaran, Tao Luo, Yan Zhou, Xiao Feng Wang, Wooseok Nam, Abhishek Sinha
-
Publication number: 20210124993Abstract: The present disclosure relates to systems, methods, and non-transitory computer readable media for training a classification neural network to classify digital images in few-shot tasks based on self-supervision and manifold mixup. For example, the disclosed systems can train a feature extractor as part of a base neural network utilizing self-supervision and manifold mixup. Indeed, the disclosed systems can apply manifold mixup regularization over a feature manifold learned via self-supervised training such as rotation training or exemplar training. Based on training the feature extractor, the disclosed systems can also train a classifier to classify digital images into novel classes not present within the base classes used to train the feature extractor.Type: ApplicationFiled: October 23, 2019Publication date: April 29, 2021Inventors: Mayank Singh, Puneet Mangla, Nupur Kumari, Balaji Krishnamurthy, Abhishek Sinha
-
Patent number: 10959232Abstract: Methods, systems, and devices for wireless communications are described. A user equipment (UE) may receive a signal identifying a first beam configuration and a second beam configuration to be used for performing a beamformed transmission of a physical uplink control channel (PUCCH) signal. The UE may determine, for a first PUCCH transmission occasion associated with the first beam configuration, that a communication metric associated with performing the beamformed transmission of the PUCCH signal using the first beam configuration fails to satisfy a threshold. The UE may perform, at a second PUCCH transmission occasion and based at least in part on the determining, the beamformed transmission of the PUCCH signal according to the second beam configuration.Type: GrantFiled: April 8, 2019Date of Patent: March 23, 2021Assignee: Qualcomm IncorporatedInventors: Xiao Feng Wang, Tao Luo, Sony Akkarakaran, Makesh Pravin John Wilson, Yan Zhou, Wooseok Nam, Abhishek Sinha, Juan Montojo
-
Publication number: 20210073267Abstract: The present disclosure relates to systems, methods, and non-transitory computer-readable media for generating tags for an object portrayed in a digital image based on predicted attributes of the object. For example, the disclosed systems can utilize interleaved neural network layers of alternating inception layers and dilated convolution layers to generate a localization feature vector. Based on the localization feature vector, the disclosed systems can generate attribute localization feature embeddings, for example, using some pooling layer such as a global average pooling layer. The disclosed systems can then apply the attribute localization feature embeddings to corresponding attribute group classifiers to generate tags based on predicted attributes. In particular, attribute group classifiers can predict attributes as associated with a query image (e.g., based on a scoring comparison with other potential attributes of an attribute group).Type: ApplicationFiled: September 9, 2019Publication date: March 11, 2021Applicant: Adobe, Inc.Inventors: Ayush Chopra, Mausoom Sarkar, Jonas Dahl, Hiresh Gupta, Balaji Krishnamurthy, Abhishek Sinha
-
Publication number: 20210042625Abstract: Methods and systems are provided for facilitating the creation and utilization of a transformation function system capable of providing network agnostic performance improvement. The transformation function system receives a representation from a task neural network. The representation can be input into a composite function neural network of the transformation function system. A learned composite function can be generated using the composite function neural network. The composite function can be specifically constructed for the task neural network based on the input representation. The learned composite function can be applied to a feature embedding of the task neural network to transform the feature embedding. Transforming the feature embedding can optimize the output of the task neural network.Type: ApplicationFiled: August 7, 2019Publication date: February 11, 2021Inventors: Ayush CHOPRA, Abhishek SINHA, Hiresh GUPTA, Mausoom SARKAR, Kumar AYUSH, Balaji KRISHNAMURTHY
-
Patent number: 10868698Abstract: A method and system are devised of moving at a NAS (1) a client (12, 13) with a MAC address, from a first VLAN to a second VLAN. A leaf is comprised of at least one intermediate L2 bridge/switch (5, 9) being connected to the NAS (1). The client (12, 13) is being connected to one (9) of the at least one intermediate L2 bridge/switches (5, 9) in the leaf. The method and system involve sending at the NAS (1) a first message downlink (31, 36) to intermediate L2 bridge/switches (5) in the leaf directly connected to the NAS.Type: GrantFiled: May 30, 2019Date of Patent: December 15, 2020Assignee: ALE USA INC.Inventors: Abhishek Sinha, Sandeep Kumar
-
Publication number: 20200382343Abstract: A method and system are devised of moving at a NAS (1) a client (12, 13) with a MAC address, from a first VLAN to a second VLAN. A leaf is comprised of at least one intermediate L2 bridge/switch (5, 9) being connected to the NAS (1). The client (12, 13) is being connected to one (9) of the at least one intermediate L2 bridge/switches (5, 9) in the leaf. The method and system involve sending at the NAS (1) a first message downlink (31, 36) to intermediate L2 bridge/switches (5) in the leaf directly connected to the NAS.Type: ApplicationFiled: May 30, 2019Publication date: December 3, 2020Inventors: Abhishek SINHA, Sandeep KUMAR