IP Library Granted Patent US 12,321,853
Granted Patent B2
US 12,321,853 · App. 17/151,132 · Granted Jun 3, 2025

Systems and methods for neural network training via local target signal augmentation

Inventors: Kurt F. Busch (Laguna Hills, CA); Jeremiah H. Holleman, III (Irvine, CA); Pieter Vorenkamp (Laguna Beach, CA); Stephen W. Bailey (Irvine, CA); David Christopher Garrett (Tustin, CA)
Assignee: SYNTIANT
G06N3/08G06N3/045G10L15/06G10L2015/088
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,321,853
App. No.
17/151,132
Granted
Jun 3, 2025
Kind
B2
Abstract

Provided herein is an integrated circuit for generating augmented training data, including a host processor configured to receive a signal stream. The integrated circuit also has a co-processor commutatively coupled to the host-processor includes a neural network with at least a first set of weights configured to identify one or more target signals from the signal stream received from the host processor. A plurality of augmentation tools are also accessible to the integrated circuit. Finally, the integrated circuit, coupled computing device or other suitable digital signal processor stores a plurality of the one or more identified target signals, and upon reaching a predetermined threshold of identified target signals, utilizes the plurality of augmentation tools to generate an extended set of target signals, and generates a second set of weights for the neural network based on the extended set of target signals generated by the augmentation tools.

Claims (41)

1. An integrated circuit for generating augmented training data, comprising:

a host processor configured to receive a signal stream and compute one or more frequency components;

a co-processor commutatively coupled to the host processor comprising a neural network with at least a first set of weights configured to identify one or more target signals from the signal stream received from the host processor;

wherein the host processor transmits the frequency components to the co-processor;

wherein the neural network of the co-processor performs inference and identifies one or more patterns from among the frequency components;

a plurality of augmentation tools;

wherein the integrated circuit further stores a plurality of the one or more identified target signals, and upon reaching a predetermined number of identified target signals, utilizes the plurality of augmentation tools to generate an extended set of target signals, and generates a second set of weights for the neural network based on the extended set of target signals generated by the augmentation tools; and

wherein one of the first or second set of weights are processed as a set of weight data that is transmitted utilizing hash encryption, or obfuscated as a subset of data within a larger set of data to be transmitted.

2. The integrated circuit of claim 1 , wherein the signal stream comprises an audio data stream.

3. The integrated circuit of claim 2 , wherein the one or more target signals comprise at least one keyword.

4. The integrated circuit of claim 3 , wherein the at least one keyword is spoken by the same user.

5. The integrated circuit of claim 4 , wherein the plurality of one or more identified target signals comprise a plurality of audio recordings of the at least one keyword or key phrase spoken by the same user.

6. The integrated circuit of claim 5 , wherein the plurality of audio recordings of the at least one keyword or key phrase spoken by the same user are of the same keyword or key phrase.

7. The integrated circuit of claim 1 , wherein the plurality of augmentation tools comprise equalization algorithms, noise generating algorithms, pitch-shifting algorithms, and time-shifting algorithms.

8. The integrated circuit of claim 1 , wherein the extended set of target signals is at least one order of magnitude greater than the plurality of the one or more identified target signals.

9. The integrated circuit of claim 1 , wherein second set of weights for the neural network are generated by utilizing the extended set of target signals as training data for the neural network.

10. The integrated circuit of claim 1 , wherein the generation of the extended set of target signals is processed on a remote device.

11. An integrated circuit for generating augmented training data, comprising:

a host processor commutatively coupled to a digital signal processor configured to receive a signal stream and compute one or more frequency components;

a co-processor commutatively coupled to the host processor comprising a neural network with at least a first set of weights configured to identify one or more target signals from the signal stream received from the host processor; and

a plurality of augmentation tools;

wherein the integrated circuit further directs the stores a plurality of the one or more identified target signals, and upon reaching a predetermined number of identified target signals, directs the digital signal processor to utilize the plurality of augmentation tools to generate an extended set of target signals, and a second set of weights for the neural network based on the extended set of target signals generated by the augmentation tools;

wherein the co-processor performs inference and identifies one or more patterns from among the frequency components; and

wherein the first and second set of weights are transmitted utilizing hash encryption, or obfuscated as a subset of data within a larger set of data to be transmitted.

12. The integrated circuit of claim 11 , wherein the signal stream comprises an audio data stream.

13. The integrated circuit of claim 12 , wherein the one or more target signals comprise at least one keyword.

14. The integrated circuit of claim 13 , wherein the at least one keyword is spoken by the same user.

15. The integrated circuit of claim 14 , wherein the plurality of one or more identified target signals comprise a plurality of audio recordings of the at least one keyword or key phrase spoken by the same user.

16. The integrated circuit of claim 15 , wherein the plurality of audio recordings of the at least one keyword or key phrase spoken by the same user are of the same keyword or key phrase.

17. The integrated circuit of claim 11 , wherein the plurality of augmentation tools comprise equalization algorithms, noise generating algorithms, pitch-shifting algorithms, and time-shifting algorithms.

18. The integrated circuit of claim 11 , wherein the extended set of target signals is at least one order of magnitude greater than the plurality of the one or more identified target signals.

19. The integrated circuit of claim 11 , wherein second set of weights for the neural network are generated by utilizing the extended set of target signals as training data for the neural network.

20. The integrated circuit of claim 11 , wherein the generation of the extended set of target signals is processed on a remote device.

21. A method for generating augmented training data, comprising:

receiving a first set of neural network weight data;

receiving a plurality of identified target signals;

computing one or more frequency components;

generating, in response to the received plurality of identified target signals exceeding a first pre-determined number, an extended set of target signals with at least one augmentation tools;

generating a second set of neural network weight data with the extended set of target signals; and

performing an inference and identifying one or more patterns from among the frequency components; and

wherein at least the first set of neural network weight data is transmitted utilizing hash encryption, or obfuscated as a subset of data within a larger set of data to be transmitted.

Assignments (2)
SECURITY INTEREST Recorded Dec 27, 2024
From: SYNTIANT CORP.; PILOT AI LABS, INC.; SYNTIANT TAIWAN LLC; SYNTIANT HOLDINGS LLC
To: OCEAN II PLO LLC
Reel/Frame 069687/0757 →
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Feb 7, 2022
From: BUSCH, KURT F.; VORENKAMP, PIETER; BAILEY, STEPHEN W.; HOLLEMAN, JEREMIAH H., III; GARRETT, DAVID CHRISTOPHER
To: SYNTIANT
Reel/Frame 058915/0250 →
Continuity (2)
Provisional Application 62962327 · Jan 17, 2020
Related Publication 20210224649A1 · Jul 22, 2021
References Cited (55)
US 6035270A · Hollier · 2000 [cited by examiner]
US 6119083A · Hollier · 2000 [cited by examiner]
US 6453284B1 · Paschall · 2002 [cited by examiner]
US 10535120B2 · Edwards · 2020 [cited by examiner]
US 10573312B1 · Thomson · 2020 [cited by examiner]
US 11328087B1 · Allen · 2022 [cited by examiner]
US 11620991B2 · Senior · 2023 [cited by examiner]
US 11741191B1 · Zilka · 2023 [cited by examiner]
US 11763829B2 · Xiao · 2023 [cited by examiner]
US 20040172238A1 · Choo · 2004 [cited by examiner]
US 20090157400A1 · Huang · 2009 [cited by examiner]
US 20100076843A1 · Ashton · 2010 [cited by examiner]
US 20110029310A1 · Jung · 2011 [cited by examiner]
US 20110125500A1 · Talwar · 2011 [cited by examiner]
US 20110137647A1 · Lee · 2011 [cited by examiner]
US 20130262096A1 · Wilhelms-Tricarico · 2013 [cited by examiner]
US 20140019756A1 · Krajec · 2014 [cited by examiner]
US 20150039299A1 · Weinstein · 2015 [cited by examiner]
US 20150127327A1 · Bacchiani · 2015 [cited by examiner]
US 20150245154A1 · Dadu · 2015 [cited by examiner]
US 20160019896A1 · Alvarez Guevara · 2016 [cited by examiner]
US 20160099007A1 · Alvarez · 2016 [cited by examiner]
US 20160111108A1 · Erdogan · 2016 [cited by examiner]
US 20160358043A1 · Mu · 2016 [cited by examiner]
US 20170032803A1 · Pandey · 2017 [cited by examiner]
US 20180366138A1 · Ramprashad · 2018 [cited by examiner]
US 20190208317A1 · Woodruff · 2019 [cited by examiner]
US 20190296910A1 · Cheung · 2019 [cited by examiner]
US 20200017117A1 · Milton · 2020 [cited by examiner]
US 20200066271A1 · Li · 2020 [cited by examiner]
US 20200125952A1 · Mosayyebpour · 2020 [cited by examiner]
US 20200126534A1 · Yoo · 2020 [cited by examiner]
US 20200150919A1 · Rand · 2020 [cited by examiner]
US 20200152330A1 · Anushiravani · 2020 [cited by examiner]
US 20200160843A1 · Shillingford · 2020 [cited by examiner]
US 20200175961A1 · Thomson · 2020 [cited by examiner]
US 20200243094A1 · Thomson · 2020 [cited by examiner]
US 20200311306A1 · Kim · 2020 [cited by examiner]
US 20200312322A1 · Cardinaux · 2020 [cited by examiner]
US 20200327884A1 · Bui · 2020 [cited by examiner]
US 20200336846A1 · Rohde · 2020 [cited by examiner]
US 20200389545A1 · Igler · 2020 [cited by examiner]
US 20210117788A1 · Muddle · 2021 [cited by examiner]
US 20210166705A1 · Chang · 2021 [cited by examiner]
US 20210193161A1 · Delcroix · 2021 [cited by examiner]
US 20210209490A1 · Casas · 2021 [cited by examiner]
US 20210248470A1 · Kaskari · 2021 [cited by examiner]
US 20210312725A1 · Milton · 2021 [cited by examiner]
US 20220068285A1 · Xiao · 2022 [cited by examiner]
US 20220108177A1 · Samek · 2022 [cited by examiner]
US 20220223161A1 · Fuchs · 2022 [cited by examiner]
US 20220414208A1 · Teranishi · 2022 [cited by examiner]
US 20230088029A1 · Guo · 2023 [cited by examiner]
Jeih-weih Hung, Hao-teng Fan and Wen-hsiang Tu, “Enhancing the magnitude spectrum of speech features for robust speech recognition.” EURASIP Journal on Advances in Signal Processing 2012,2012:189. (Year: 2012). [cited by examiner]
PCT International Search Report and Written Opinion, PCT Application No. PCT/US2021/013768, mailed Mar. 31, 2021. [cited by applicant]