IP Library Granted Patent US 11,675,876
Granted Patent B2
US 11,675,876 · App. 17/082,060 · Granted Jun 13, 2023

Training robust machine learning models

Inventor: Amod Jog (Mountain View, CA)
Assignee: International Business Machines Corporation
G06F18/2135G06T7/0012G06T2207/20081G06T2207/20084G06V2201/03
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 11,675,876
App. No.
17/082,060
Granted
Jun 13, 2023
Kind
B2
Abstract

Training a robust machine learning model by mapping an input data set to a first feature space, applying a transformation to the first feature space, yielding a second feature space, and training a dense model using the first feature space, and the second feature space.

Claims (32)

1. A computer implemented method for training a robust machine learning model, the method comprising:

mapping, by one or more computer processors, an input data set to a first feature space;

applying, by the one or more computer processors, a linear transformation to the first feature space, yielding a second feature space; and

training, by the one or more computer processors, a dense model using the first feature space and the second feature space.

2. The computer implemented method according to claim 1 , wherein the dense model comprises a classifier.

3. The computer implemented method according to claim 1 , wherein mapping the input data to a feature space comprises mapping the input data with a convolutional neural network.

4. The computer implemented method according to claim 1 , wherein the dense model classifies image features.

5. The computer implemented method according to claim 1 , wherein the input data set comprises medical diagnostic images.

6. The computer implemented method according to claim 1 , wherein applying a transformation to the first feature space comprises applying a linear transformation matrix to the first feature space.

7. The computer implemented method according to claim 1 , further comprising providing, by the one or more computer processors, a trained model configured to classify input data.

8. A computer program product for training a robust machine learning model, the computer program product comprising one or more computer readable storage devices and collectively stored program instructions on the one or more computer readable storage devices, the stored program instructions comprising:

program instructions to map an input data set to a first feature space;

program instructions to apply a linear transformation to the first feature space, yielding a second feature space; and

program instructions to train a dense model using the first feature space and the second feature space.

9. The computer program product according to claim 8 , wherein the dense model comprises a classifier.

10. The computer program product according to claim 8 , the stored program instructions further comprising program instructions to map the input data to a feature space with a convolutional neural network.

11. The computer program product according to claim 8 , the stored program instructions further comprising program instructions to classify image features.

12. The computer program product according to claim 8 , wherein the input data comprises medical diagnostic images.

13. The computer program product according to claim 8 , wherein the stored program instructions to apply a transformation to the first feature space comprise program instructions to apply a linear transformation matrix to the first feature space.

14. The computer program product according to claim 8 , the stored program instructions further comprising program instructions to provide a trained model configured to classify input data.

15. A computer system for training a robust machine learning model, the computer system comprising:

one or more computer processors;

one or more computer readable storage devices; and

stored program instructions on the one or more computer readable storage devices for execution by the one or more computer processors, the stored program instructions comprising:

program instructions to map an input data set to a first feature space;

program instructions to apply a linear transformation to the first feature space, yielding a second feature space; and

program instructions to train a dense model using the first feature space and the second feature space.

16. The computer system according to claim 15 , the stored program instructions further comprising program instructions to map the input data to a feature space with a convolutional neural network.

17. The computer system according to claim 15 , the stored program instructions further comprising program instructions to classify image features.

18. The computer system according to claim 15 , wherein the input data comprises medical diagnostic images.

19. The computer system according to claim 15 , wherein the stored program instructions to apply a transformation to the first feature space comprise program instructions to apply a linear transformation matrix to the first feature space.

20. The computer system according to claim 15 , the stored program instructions further comprising program instructions to provide a trained model configured to classify input data.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Oct 28, 2020
From: JOG, AMOD
To: INTERNATIONAL BUSINESS MACHINES CORPORATION
Reel/Frame 054188/0057 →
Continuity (1)
Related Publication 20220129697A1 · Apr 28, 2022