Patents by Inventor Sandeep Joshi
Sandeep Joshi 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: 11960433Abstract: Apparatuses, systems, and techniques to route data transfers between hardware devices. In at least one embodiment, a path over which to transfer data from a first hardware component of a computer system to a second hardware component of a computer system is determined based, at least in part, on one or more characteristics of different paths usable to transfer the data.Type: GrantFiled: September 23, 2021Date of Patent: April 16, 2024Assignee: NVIDIA Technologies, Inc.Inventors: Kiran Kumar Modukuri, Christopher J. Newburn, Saptarshi Sen, Akilesh Kailash, Sandeep Joshi
-
Patent number: 11921787Abstract: Systems and methods are provided for organizing communications of multiple communication services. Specifically, techniques are described for organizing data according to detected identities represented aggregate accounts across multiple communication services, as opposed to merely on the basis of accounts on such services. An identity-aware data management system is disclosed that can generate, on the basis of communications of a first party among multiple communication services, a reduced connection graph of the first party. The reduced connection graph of the first party can reflect second parties with whom the first party has communicated, each second party representing an aggregation of multiple accounts of a second party across communication services. The identity-aware data management system may then enable the first party to search, filter, or organize communications according to the reduced connection graph.Type: GrantFiled: July 14, 2022Date of Patent: March 5, 2024Assignee: NEEDL.AI INC.Inventors: Sandeep Joshi, Raghul Ravi, Vikram Srinivasan, Kuntal Shah
-
Patent number: 11900700Abstract: Systems, methods, and computer-readable media are disclosed for language-agnostic subtitle drift detection and correction. A method may include determining subtitles and/or captions from media content (e.g., videos), the subtitles and/or captions corresponding to dialog in the media content. The subtitles may be broken up into segments which may be analyzed to determine a likelihood of drift (e.g., a likelihood that the subtitles are out of synchronization with the dialog in the media content) for each segment. For segments with a high likelihood of drift, the subtitles may be incrementally adjusted to determine an adjustment that eliminates and/or reduces the amount of drift, and the drift in the segment may be corrected based on the drift amount detected. A linear regression model and/or human blocks determined by human operators may be used to otherwise optimize drift correction.Type: GrantFiled: February 27, 2023Date of Patent: February 13, 2024Assignee: Amazon Technologies, Inc.Inventors: Tamojit Chatterjee, Mayank Sharma, Muhammad Raffay Hamid, Sandeep Joshi
-
Publication number: 20240020340Abstract: Systems and methods are provided for organizing communications of multiple communication services. Specifically, techniques are described for organizing data according to detected identities represented aggregate accounts across multiple communication services, as opposed to merely on the basis of accounts on such services. An identity-aware data management system is disclosed that can generate, on the basis of communications of a first party among multiple communication services, a reduced connection graph of the first party. The reduced connection graph of the first party can reflect second parties with whom the first party has communicated, each second party representing an aggregation of multiple accounts of a second party across communication services. The identity-aware data management system may then enable the first party to search, filter, or organize communications according to the reduced connection graph.Type: ApplicationFiled: July 14, 2022Publication date: January 18, 2024Inventors: Sandeep Joshi, Raghul Ravi, Vikram Srinivasan, Kuntal Shah
-
Patent number: 11869537Abstract: Systems, methods, and computer-readable media are disclosed for systems and methods for language agnostic automated voice activity detection. Example methods may include determining an audio file associated with video content, generating audio segments using the audio file, the audio segments including a first segment and a second segment, and determining that the first segment includes first voice activity. Methods may include determining that the second segment comprises second voice activity, determining that voice activity is present between a first timestamp associated with the first segment and a second timestamp associated with the second segment, and generating text data representing the voice activity that is present between the first timestamp and the second timestamp.Type: GrantFiled: November 10, 2021Date of Patent: January 9, 2024Assignee: Amazon Technologies, Inc.Inventors: Mayank Sharma, Sandeep Joshi, Muhammad Raffay Hamid
-
Publication number: 20230282006Abstract: Systems, methods, and computer-readable media are disclosed for language-agnostic subtitle drift detection and correction. A method may include determining subtitles and/or captions from media content (e.g., videos), the subtitles and/or captions corresponding to dialog in the media content. The subtitles may be broken up into segments which may be analyzed to determine a likelihood of drift (e.g., a likelihood that the subtitles are out of synchronization with the dialog in the media content) for each segment. For segments with a high likelihood of drift, the subtitles may be incrementally adjusted to determine an adjustment that eliminates and/or reduces the amount of drift, and the drift in the segment may be corrected based on the drift amount detected. A linear regression model and/or human blocks determined by human operators may be used to otherwise optimize drift correction.Type: ApplicationFiled: February 27, 2023Publication date: September 7, 2023Applicant: Amazon Technologies, Inc.Inventors: Tamojit Chatterjee, Mayank Sharma, Muhammad Raffay Hamid, Sandeep Joshi
-
Patent number: 11625928Abstract: Systems, methods, and computer-readable media are disclosed for language-agnostic subtitle drift detection and correction. A method may include determining subtitles and/or captions from media content (e.g., videos), the subtitles and/or captions corresponding to dialog in the media content. The subtitles may be broken up into segments which may be analyzed to determine a likelihood of drift (e.g., a likelihood that the subtitles are out of synchronization with the dialog in the media content) for each segment. For segments with a high likelihood of drift, the subtitles may be incrementally adjusted to determine an adjustment that eliminates and/or reduces the amount of drift and the drift in the segment may be corrected based on the drift amount detected. A linear regression model and/or human blocks determined by human operators may be used to otherwise optimize drift correction.Type: GrantFiled: September 1, 2020Date of Patent: April 11, 2023Assignee: AMAZON TECHNOLOGIES, INC.Inventors: Tamojit Chatterjee, Mayank Sharma, Muhammad Raffay Hamid, Sandeep Joshi
-
Publication number: 20220286325Abstract: In an embodiment, a method of a base station (BS) for managing an ISI in a cellular network is disclosed. The method includes receiving at least one UE-Capability information element from a UE comprising a list of CP lengths and a list of SCSs, determining a plurality of parameters associated with the UE based on the list of CP lengths and the list of SCSs, calculating at least one of a first custom CP length and a first SCS based on the plurality of parameters from the list of CP lengths and the list of SCSs, and transmitting, to the UE, a response message indicating that at least one of the first custom CP length and the first SCS is selected for managing the ISI.Type: ApplicationFiled: March 4, 2022Publication date: September 8, 2022Inventors: Anusha GUNTURU, Ashok Kumar Reddy CHAVVA, Sandeep JOSHI
-
Publication number: 20220012207Abstract: Apparatuses, systems, and techniques to route data transfers between hardware devices. In at least one embodiment, a path over which to transfer data from a first hardware component of a computer system to a second hardware component of a computer system is determined based, at least in part, on one or more characteristics of different paths usable to transfer the data.Type: ApplicationFiled: September 23, 2021Publication date: January 13, 2022Inventors: Kiran Kumar Modukuri, Christopher J. Newburn, Saptarshi Sen, Akilesh Kailash, Sandeep Joshi
-
Patent number: 11205445Abstract: Systems, methods, and computer-readable media are disclosed for systems and methods for language agnostic automated voice activity detection. Example methods may include determining an audio file associated with video content, generating a number of audio segments using the audio file, the plurality of audio segments including a first segment and a second segment, where the first segment and the second segment are consecutive segments. Example methods may include determining, using a Gated Recurrent Unit neural network, that the first segment includes first voice activity, determining, using the Gated Recurrent Unit neural network, that the second segment includes second voice activity, and determining that voice activity is present between a first timestamp associated with the first segment and a second timestamp associated with the second segment.Type: GrantFiled: June 10, 2019Date of Patent: December 21, 2021Assignee: Amazon Technologies, Inc.Inventors: Mayank Sharma, Sandeep Joshi, Muhammad Raffay Hamid
-
Patent number: 11132326Abstract: Apparatuses, systems, and techniques to route data transfers between hardware devices. In at least one embodiment, a path over which to transfer data from a first hardware component of a computer system to a second hardware component of a computer system is determined based, at least in part, on one or more characteristics of different paths usable to transfer the data.Type: GrantFiled: March 11, 2020Date of Patent: September 28, 2021Assignee: NVIDIA CorporationInventors: Kiran Kumar Modukuri, Christopher J. Newburn, Saptarshi Sen, Akilesh Kailash, Sandeep Joshi
-
Publication number: 20210286752Abstract: Apparatuses, systems, and techniques to route data transfers between hardware devices. In at least one embodiment, a path over which to transfer data from a first hardware component of a computer system to a second hardware component of a computer system is determined based, at least in part, on one or more characteristics of different paths usable to transfer the data.Type: ApplicationFiled: March 11, 2020Publication date: September 16, 2021Inventors: Kiran Kumar Modukuri, Christopher J. Newburn, Saptarshi Sen, Akilesh Kailash, Sandeep Joshi
-
Patent number: 10945041Abstract: Devices, systems, and methods are provided for language-agnostic subtitle drift detection and localization. A method may include extracting audio from video, dividing the audio into overlapping blocks, and determining the probabilities of overlapping portions of the blocks, the probabilities indicating a presence of voice data represented by the audio in the blocks. The method may generate machine blocks using overlapping portions of blocks where voice data is present, and may map the machine blocks to corresponding blocks indicating that subtitles are available for the video. For mapped blocks, the method may include determining features such as when subtitles are available without voice audio, when voice audio is available without subtitles, and when voice audio and subtitles both are available. Using the features, the method may include determining the probability that the video includes subtitle drift, and the method may include analyzing the video to localize where the subtitle drift occurs.Type: GrantFiled: June 2, 2020Date of Patent: March 9, 2021Assignee: Amazon Technologies, Inc.Inventors: Tamojit Chatterjee, Mayank Sharma, Muhammad Raffay Hamid, Sandeep Joshi
-
Patent number: 10920194Abstract: Hair follicle bulge region/LLP region CD34(+) MeSCs can be isolated from mammalian skin bearing hair follicles. These cells are multipotent and retain the ability to differentiate into cells of neural crest lineage, including glia-like cells that express the glial marker Gfap, and are able to express myelin basic protein, and to remyelinate naked (unmyelinated or demyelinated) neuronal processes with a functional, dense myelin sheath. These cells of neural crest lineage can be used to produce a dense myelin sheath on neurons which lack myelin due to genetic defect, trauma, toxin, infection, or disease process. Therefore, embodiments of the invention provide methods for preparing such cells, the cells themselves and compositions containing the cells, as well as methods for using the cells.Type: GrantFiled: November 3, 2017Date of Patent: February 16, 2021Assignees: UNIVERSITY OF MARYLAND, BALTIMORE, THE UNITED STATES OF AMERICA AS REPRESENTED BY THE DEPARTMENT OF VETERANS AFFAIRSInventors: Thomas Hornyak, Sandeep Joshi
-
Publication number: 20180057790Abstract: Hair follicle bulge region/LLP region CD34(+) MeSCs can be isolated from mammalian skin bearing hair follicles. These cells are multipotent and retain the ability to differentiate into cells of neural crest lineage, including glia-like cells that express the glial marker Gfap, and are able to express myelin basic protein, and to remyelinate naked (unmyelinated or demyelinated) neuronal processes with a functional, dense myelin sheath. These cells of neural crest lineage can be used to produce a dense myelin sheath on neurons which lack myelin due to genetic defect, trauma, toxin, infection, or disease process. Therefore, embodiments of the invention provide methods for preparing such cells, the cells themselves and compositions containing the cells, as well as methods for using the cells.Type: ApplicationFiled: November 3, 2017Publication date: March 1, 2018Inventors: Thomas Hornyak, Sandeep Joshi
-
Patent number: 9740701Abstract: Implementations disclosed herein provide for cauterizing portions of a snapshot. A file and/or directory or a set of files and directories of a snapshot can be identified via a user request or an automated process for cauterization. When cauterized, if a previous snapshot does not contain a ditto entry to the portion of the current snapshot that is requested to be cauterized, the BADDR associated with the cauterized portion of the snapshot can be freed for other storage use. In some implementations, metadata associated with the data that is cauterized can be retained for other uses.Type: GrantFiled: June 30, 2014Date of Patent: August 22, 2017Assignee: EMC IP Holding CompanyInventors: Mohd Fahadullah, Kamal J Charan, Vikrama Reddy, Prasad Krishnan, Sandeep Joshi
-
Patent number: 9501487Abstract: Implementations are provided herein for incremental backup using a change tree. A change tree is a database file for tracking file system changes of a specific dataset, such as a directory of a file system. A backup process can create and/or initiate a change tree when creating a backup snapshot of a dataset. After the change tree is created, all file system changes inside the dataset can be tracked. The next incremental backup can then take advantage of the change tree to backup changes without traversing the entire dataset. Thus, incremental backups can be more efficient and are better able to keep up with tape streaming speed.Type: GrantFiled: June 30, 2014Date of Patent: November 22, 2016Assignee: EMC CorporationInventors: Hsing Yuan, Suchit Kaura, Hymanand Nellore, Joseph Jobi, Sandeep Joshi, Moiz Haidry, Ravi Venkata Batchu, Narain Coimbatore Ramadass
-
Patent number: 9384200Abstract: The subject disclosure is directed towards the parallel backing up of a file system. A file system tree structure is walked by parallel workers that write file system data to data storage devices in parallel streams. Work assigned to one worker may be split to an idle worker to provide parallel operation. Relationship data that maintains relationships between the streams is persisted (e.g., via reference data inserted into the streams) so that a restore mechanism is able to stitch the streams together back into the file system data. Also persisted is location data that maintains storage locations of the streams. A file's data also may be written into parallel streams.Type: GrantFiled: December 21, 2012Date of Patent: July 5, 2016Assignee: EMC CorporationInventors: Ravi V. Batchu, Suchit Kaura, Hymanand Nellore, Hsing Yuan, Jeff Miller, Sandeep Joshi
-
Patent number: 9348674Abstract: A first request to execute a first task is received from a first module in a first address space and by a second module in a second address space. The first task is placed into a task queue for execution in the second address space. Pending responses not yet returned to the first module that are results of execution for other tasks in the second address space are extracted by the second module from a response queue. Requests for the other tasks were previously sent by the first module to the second module for execution in the second address space. The pending responses are compounded. The pending responses and a return value for acknowledgement the first request to execute the first task are combined, by the second module into a combined communication. The combined communication is transmitted by the second module to the first module in the first address space.Type: GrantFiled: July 25, 2014Date of Patent: May 24, 2016Assignee: NetApp, Inc.Inventors: Sandeep Joshi, Ankit Mathur, Sudip Kumar Panda
-
Publication number: 20160060597Abstract: Hair follicle bulge region/LLP region CD34(+) MeSCs can be isolated from mammalian skin bearing hair follicles. These cells are multipotent and retain the ability to differentiate into cells of neural crest lineage, including glia-like cells that express the glial marker Gfap, and are able to express myelin basic protein, and to remyelinate naked (unmyelinated or demyelinated) neuronal processes with a functional, dense myelin sheath. These cells of neural crest lineage can be used to produce a dense myelin sheath on neurons which lack myelin due to genetic defect, trauma, toxin, infection, or disease process. Therefore, embodiments of the invention provide methods for preparing such cells, the cells themselves and compositions containing the cells, as well as methods for using the cells.Type: ApplicationFiled: August 27, 2015Publication date: March 3, 2016Inventors: Thomas Hornyak, Sandeep Joshi