Title:
NEURAL NETWORK WITH IMPROVED PERFORMANCE, USING AUTOMATICALLY UNCOVERED FAILURE CASES
Document Type and Number:
Japanese Patent JP2023109726
Kind Code:
A
Abstract:
To provide a method, system and apparatus, including a computer program encoded on a computer storage medium, for adjusting a target neural network using automatically generated test cases before deployment of the target neural network in a deployment environment.SOLUTION: One of methods may include the steps of: generating a plurality of test inputs by using a test case generation neural network; processing the plurality of test inputs using a target neural network to generate one or more test outputs for each test input; and identifying, from the one or more test outputs generated by the target neural network for each test input, failing test inputs that result in generation of test outputs by the target neural network that fail one or more criteria.SELECTED DRAWING: Figure 3
Inventors:
ETHAN JOSEAN PEREZ
SAFFRON SHAN HUANG
NATHANIEL JOHN MCALEESE PARK
GEOFFREY IRVING
SAFFRON SHAN HUANG
NATHANIEL JOHN MCALEESE PARK
GEOFFREY IRVING
Application Number:
JP2023010303A
Publication Date:
August 08, 2023
Filing Date:
January 26, 2023
Export Citation:
Assignee:
DEEPMIND TECH LITD
International Classes:
G06N3/045; G06N3/0475; G06N3/08
Attorney, Agent or Firm:
Yasuhiko Murayama
Shinya Mihiro
Tatsuhiko Abe
Shinya Mihiro
Tatsuhiko Abe
Previous Patent: Footwear product sole structure
Next Patent: Bond coat containing coarse oxygen getter particles
Next Patent: Bond coat containing coarse oxygen getter particles