Hyun Suk Park

Author:

Hyun Ok Park

ISBN #:

0822336146

Author:

Hyun Ok Park

ISBN #:

0822336251

US Patent:

7572904, Aug 11, 2009

Filed:

Mar 26, 2004

Appl. No.:

10/811508

Inventors:

Xing Cheng - Sunnyvale CA, US
Hyun Jung Park - Sunnyvale CA, US
Hong Jin - Cupertino CA, US

Assignee:

MedImmune, LLC - Gaithersburg MD

International Classification:

C07H 21/00
A61K 39/155

US Classification:

536 2372, 4242111

Abstract:

The complete polynucleotide sequence of the human respiratory syncytial virus subgroup B strain 9320 genome is provided. Proteins encoded by this polynucleotide sequence are also provided. Isolated or recombinant RSV (e. g. , attenuated recombinant RSV), nucleic acids, and polypeptides, e. g. , comprising mutations in the attachment protein G, are also provided, as are immunogenic compositions comprising such isolated or recombinant RSV, nucleic acids, and polypeptides. Related methods are also described.

US Patent:

8163530, Apr 24, 2012

Filed:

Jul 23, 2009

Appl. No.:

12/508075

Inventors:

Xing Cheng - Sunnyvale CA, US
Hyun Jung Park - Sunnyvale CA, US
Hong Jin - Cupertino CA, US

Assignee:

MedImmune, LLC - Gaithersburg MD

International Classification:

C12N 7/00
C12N 7/01
C12N 7/04
A61K 39/155

US Classification:

4352351, 435236, 4242111, 536 2372

Abstract:

The complete polynucleotide sequence of the human respiratory syncytial virus subgroup B strain 9320 genome is provided. Proteins encoded by this polynucleotide sequence are also provided. Isolated or recombinant RSV (e. g. , attenuated recombinant RSV), nucleic acids, and polypeptides, e. g. , comprising mutations in the attachment protein G, are also provided, as are immunogenic compositions comprising such isolated or recombinant RSV, nucleic acids, and polypeptides. Related methods are also described.

US Patent:

20120263744, Oct 18, 2012

Filed:

Apr 17, 2012

Appl. No.:

13/449087

Inventors:

Xing Cheng - Sunnyvale CA, US
Hyun Jung Park - Sunnyvale CA, US
Hong Jin - Cupertino CA, US

Assignee:

MedImmune, LLC - Gaithersburg MD

International Classification:

A61K 39/155
A61P 31/14

US Classification:

4241861, 4242111

Abstract:

The complete polynucleotide sequence of the human respiratory syncytial virus subgroup B strain 9320 genome is provided. Proteins encoded by this polynucleotide sequence are also provided. Isolated or recombinant RSV (e.g., attenuated recombinant RSV), nucleic acids, and polypeptides, e.g., comprising mutations in the attachment protein G, are also provided, as are immunogenic compositions comprising such isolated or recombinant RSV, nucleic acids, and polypeptides. Related methods are also described.

US Patent:

20210312913, Oct 7, 2021

Filed:

Jun 15, 2021

Appl. No.:

17/348422

Inventors:

- Mountain View CA, US
Hyun Jin Park - Mountain View CA, US
Patrick Violette - Mountain View CA, US

Assignee:

Google LLC - Mountain View CA

International Classification:

G10L 15/16
G10L 15/02
G10L 15/06
G10L 15/22

Abstract:

A method for training hotword detection includes receiving a training input audio sequence including a sequence of input frames that define a hotword that initiates a wake-up process on a device. The method also includes feeding the training input audio sequence into an encoder and a decoder of a memorized neural network. Each of the encoder and the decoder of the memorized neural network include sequentially-stacked single value decomposition filter (SVDF) layers. The method further includes generating a logit at each of the encoder and the decoder based on the training input audio sequence. For each of the encoder and the decoder, the method includes smoothing each respective logit generated from the training input audio sequence, determining a max pooling loss from a probability distribution based on each respective logit, and optimizing the encoder and the decoder based on all max pooling losses associated with the training input audio sequence.

US Patent:

20210142790, May 13, 2021

Filed:

Jan 21, 2021

Appl. No.:

17/155068

Inventors:

- Mountain View CA, US
Hyun Jin Park - Mountain View CA, US

Assignee:

Google LLC - Mountain View CA

International Classification:

G10L 15/16
G10L 15/02
G10L 15/06
G10L 15/22

Abstract:

A method for detecting a hotword includes receiving a sequence of input frames that characterize streaming audio captured by a user device and generating a probability score indicating a presence of a hotword in the streaming audio using a memorized neural network. The network includes sequentially-stacked single value decomposition filter (SVDF) layers and each SVDF layer includes at least one neuron. Each neuron includes a respective memory component, a first stage configured to perform filtering on audio features of each input frame individually and output to the memory component, and a second stage configured to perform filtering on all the filtered audio features residing in the respective memory component. The method also includes determining whether the probability score satisfies a hotword detection threshold and initiating a wake-up process on the user device for processing additional terms.

US Patent:

20200126537, Apr 23, 2020

Filed:

Dec 10, 2019

Appl. No.:

16/709191

Inventors:

- Mountain View CA, US
Hyun Jin Park - Mountain View CA, US
Patrick Violette - Mountain View CA, US

Assignee:

Google LLC - Mountain View CA

International Classification:

G10L 15/16
G10L 15/06
G10L 15/02
G10L 15/22

Abstract:

A method for training hotword detection includes receiving a training input audio sequence including a sequence of input frames that define a hotword that initiates a wake-up process on a device. The method also includes feeding the training input audio sequence into an encoder and a decoder of a memorized neural network. Each of the encoder and the decoder of the memorized neural network include sequentially-stacked single value decomposition filter (SVDF) layers. The method further includes generating a logit at each of the encoder and the decoder based on the training input audio sequence. For each of the encoder and the decoder, the method includes smoothing each respective logit generated from the training input audio sequence, determining a max pooling loss from a probability distribution based on each respective logit, and optimizing the encoder and the decoder based on all max pooling losses associated with the training input audio sequence.

US Patent:

20200020322, Jan 16, 2020

Filed:

Jun 13, 2019

Appl. No.:

16/439897

Inventors:

- Mountain View CA, US
Hyun Jin Park - Mountain View CA, US

Assignee:

Google LLC - Mountain View CA

International Classification:

G10L 15/16
G10L 15/22
G10L 15/02
G10L 15/06

Abstract:

A method for detecting a hotword includes receiving a sequence of input frames that characterize streaming audio captured by a user device and generating a probability score indicating a presence of a hotword in the streaming audio using a memorized neural network. The network includes sequentially-stacked single value decomposition filter (SVDF) layers and each SVDF layer includes at least one neuron. Each neuron includes a respective memory component, a first stage configured to perform filtering on audio features of each input frame individually and output to the memory component, and a second stage configured to perform filtering on all the filtered audio features residing in the respective memory component. The method also includes determining whether the probability score satisfies a hotword detection threshold and initiating a wake-up process on the user device for processing additional terms.