Vivek Kumar

US Patent:

20220335925, Oct 20, 2022

Filed:

Aug 18, 2020

Appl. No.:

17/636851

Inventors:

- San Francisco, US
Xiaoyu LIU - San Mateo CA, US
Michael Getty HORGAN - San Francisco CA, US
Vivek Kumar - Foster City CA, US

Assignee:

DOLBY LABORATORIES LICENSING CORPORATION - San Francisco CA

International Classification:

G10L 13/033
G10L 13/047

Abstract:

Novel methods and systems for adapting a voice cloning synthesizer for a new speaker using real speech data are disclosed. Utterances from one or more target speakers are parameterized and are used to initialize an embedding vector for use with a voice synthesizer, by means of clustering the utterance data and determining the centroid of the data, using a speaker identification neural network, and/or by finding the closest stored embedded vector to the utterance data.

US Patent:

20220311083, Sep 29, 2022

Filed:

Mar 24, 2021

Appl. No.:

17/211209

Inventors:

- Dearborn MI, US
Sumanth Reddy Dadam - New Hudson MI, US
Vivek Kumar - Troy MI, US
Vinod Ravi - Canton MI, US

International Classification:

H01M 50/317
H01M 10/52
H01M 50/358
B60K 1/04

Abstract:

A traction battery venting system includes, among other things, an engine exhaust system that discharges engine exhaust gas to an area outside an electrified vehicle. The engine exhaust gas is emitted from an internal combustion engine of the electrified vehicle. The system further includes a sound deadening device of the engine exhaust system, a traction battery, and a battery vent system that communicates battery vent byproducts to the engine exhaust system at a position that is downstream from the sound deadening device. The battery vent byproducts are emitted from the traction battery of the electrified vehicle.

US Patent:

20220298941, Sep 22, 2022

Filed:

Mar 16, 2021

Appl. No.:

17/203279

Inventors:

- Dearborn MI, US
Vinod Ravi - Canton MI, US
Di Zhu - Canton MI, US
Vivek Kumar - Troy MI, US

International Classification:

F01N 1/16
G10K 11/16

Abstract:

Methods and systems are provided for adjusting a position of an exhaust valve that regulates engine exhaust noise. In one example, the position of the exhaust valve is adjusted according to a distance between a first vehicle and a second vehicle. The exhaust valve may be adjusted to reduce an exhaust noise level when the distance between the first vehicle and the second vehicle is less than a threshold distance.

US Patent:

20220259992, Aug 18, 2022

Filed:

Feb 12, 2021

Appl. No.:

17/175286

Inventors:

- Dearborn MI, US
Vivek Kumar - Troy MI, US
Douglas Martin - Canton MI, US
Vinod Ravi - Canton MI, US

International Classification:

F01M 5/00
F01M 5/04
F01M 5/02

Abstract:

Methods and systems are provided for controlling engine oil dilution in automotive vehicles. The method comprises heating coolant by operating an exhaust gas heat recovery device and circulating the heated coolant via a first coolant loop between the exhaust gas heat recovery device and a heat exchanger; and heating coolant by operating an exhaust gas recirculation cooler and circulating the heated coolant via a second coolant loop between the exhaust gas recirculation cooler and the heat exchanger, wherein the heated coolant from both the exhaust gas heat recovery device and the exhaust gas recirculation cooler mix at the heat exchanger and allows heating of an engine oil via the heat exchanger. In one example, the method prevents excessive accumulation of water and/or fuel in the engine oil.

US Patent:

20220253635, Aug 11, 2022

Filed:

Mar 3, 2021

Appl. No.:

17/190663

Inventors:

- Detroit MI, US
Sycamore Zhang - Shanghai, CN
Shuqing Zeng - Sterling Heights MI, US
Kamran Ali - Troy MI, US
Vivek Vijaya Kumar - Shelby Township MI, US
Donald K. Grimm - Utica MI, US

International Classification:

G06K 9/00
G06K 9/62
G06K 9/32
B60W 60/00
B60W 30/08
G06K 9/68

Abstract:

An attention-driven streaming system includes an adaptive-streaming module and a transceiver. The adaptive-streaming module includes filters, a compression module, an attention-driven strategy module and a fusion module. The filters filter sensor data received from sensors of a vehicle. The compression module compresses the filtered sensor data to generate compressed data. The attention-driven strategy module generates feedforward information based on a state of the vehicle and a state of an environment of the vehicle to adjust a region of interest. The fusion module generates an adaptive streaming strategy to adaptively adjust operations of each of the filters. The transceiver streams the compressed data to at least one of an edge computing device or a cloud-based network device and in response receive feedback information and pipeline monitoring information. The fusion module generates the adaptive streaming strategy based on the feedforward information, the feedback information and the pipeline monitoring information.

US Patent:

20220245506, Aug 4, 2022

Filed:

Jan 30, 2021

Appl. No.:

17/163396

Inventors:

- Bentonville AR, US
Adrian Sonjong Yi - San Francisco CA, US
Peng Yang - San Jose CA, US
Valeriy Pelyushenko - San Jose CA, US
Haibo Yan - San Bruno CA, US
Vivek Kumar - Bellevue WA, US
Jayanth Korlimarla - Sunnyvale CA, US
Rajesh Garigipati - Dublin CA, US

International Classification:

G06N 20/00
G06F 16/903

Abstract:

Systems and methods for selecting items from a pool of items based on comparisons of composite embeddings are disclosed. A first initial embedding is obtained from a first database for each item in a pool of items and a second initial embedding is obtained from a second database for each item in the pool of items. The first initial embedding is generated using a first embedding model and the second initial embedding is generated using a second embedding model. A first composite embedding is generated for each item in the pool of items comprising the first initial embedding and the second initial embedding; and compare the first composite embedding for each item in the pool of items to a first anchor embedding, wherein the first anchor embedding comprises a first initial anchor embedding generated using the first embedding model and a second initial anchor embedding generated using the second embedding model.

US Patent:

20210142658, May 13, 2021

Filed:

Nov 11, 2019

Appl. No.:

16/679791

Inventors:

- Detroit MI, US
Fan Bai - Ann Arbor MI, US
Bo Yu - Troy MI, US
Vivek Vijaya Kumar - Shelby Township MI, US

International Classification:

G08G 1/01
G06K 9/00
G08G 1/07

Abstract:

Systems and methods to perform traffic signal management for operation of autonomous vehicles involve obtaining vehicle data from two or more vehicles at an intersection with one or more traffic lights. The vehicle data includes vehicle location, vehicle speed, and image information or images. A method includes determining at least one of three types of information about the one or more traffic lights based on the vehicle data. The three types of information include a location of the one or more traffic lights, a signal phase and timing (SPaT) of the one or more traffic lights, and a lane correspondence of the one or more traffic lights. The method also includes providing the at least one of the three types of information about the one or more traffic lights for the operation of autonomous vehicles.

US Patent:

20210082444, Mar 18, 2021

Filed:

Apr 10, 2019

Appl. No.:

17/046284

Inventors:

- San Francisco CA, US
Grant A. DAVIDSON - Burlingame CA, US
Chih-Wei WU - San Francisco CA, US
Vivek KUMAR - Foster City CA, US

Assignee:

Dolby Laboratories Licensing Corporation - San Francisco CA

International Classification:

G10L 19/022
G06F 3/16
G06N 3/08
G06N 3/04

Abstract:

Computer-implemented methods for training a neural network, as well as for implementing audio encoders and decoders via trained neural networks, are provided. The neural network may receive an input audio signal, generate an encoded audio signal and decode the encoded audio signal. A loss function generating module may receive the decoded audio signal and a ground truth audio signal, and may generate a loss function value corresponding to the decoded audio signal. Generating the loss function value may involve applying a psychoacoustic model. The neural network may be trained based on the loss function value. The training may involve updating at least one weight of the neural network.