17994429. AUTOMATIC DATA FABRICATION BY COMBINING GENERATIVE ADVERSARIAL NETWORKS simplified abstract (INTERNATIONAL BUSINESS MACHINES CORPORATION)
Contents
- 1 AUTOMATIC DATA FABRICATION BY COMBINING GENERATIVE ADVERSARIAL NETWORKS
- 1.1 Organization Name
- 1.2 Inventor(s)
- 1.3 AUTOMATIC DATA FABRICATION BY COMBINING GENERATIVE ADVERSARIAL NETWORKS - A simplified explanation of the abstract
- 1.4 Simplified Explanation
- 1.5 Potential Applications
- 1.6 Problems Solved
- 1.7 Benefits
- 1.8 Potential Commercial Applications
- 1.9 Possible Prior Art
- 1.10 Unanswered Questions
- 1.11 Original Abstract Submitted
AUTOMATIC DATA FABRICATION BY COMBINING GENERATIVE ADVERSARIAL NETWORKS
Organization Name
INTERNATIONAL BUSINESS MACHINES CORPORATION
Inventor(s)
Omer Yehuda Boehm of Haifa (IL)
AUTOMATIC DATA FABRICATION BY COMBINING GENERATIVE ADVERSARIAL NETWORKS - A simplified explanation of the abstract
This abstract first appeared for US patent application 17994429 titled 'AUTOMATIC DATA FABRICATION BY COMBINING GENERATIVE ADVERSARIAL NETWORKS
Simplified Explanation
The abstract describes a computer-implemented method that involves training two Generative Adversarial Networks (GANs) based on original and fabricated structured data, combining them into a combined GAN, training the combined GAN, and using it to generate new fabricated data that imitate characteristics of the original data and adhere to user-defined constraints.
- Training two GANs: One based on original structured data and the other on fabricated structured data that adhere to user-defined constraints.
- Combining the two GANs into a single combined GAN.
- Training the combined GAN.
- Generating new fabricated data that imitate characteristics of the original data and adhere to user-defined constraints.
Potential Applications
This technology could be applied in various fields such as data generation, data augmentation, and synthetic data creation for training machine learning models.
Problems Solved
This technology addresses the challenge of generating new data that closely resembles original data while adhering to specific constraints set by the user.
Benefits
The technology allows for the creation of diverse and realistic data that can be used to enhance the performance of machine learning models and improve data analysis processes.
Potential Commercial Applications
Potential commercial applications of this technology include data science, artificial intelligence, and machine learning industries for tasks such as image generation, text generation, and anomaly detection.
Possible Prior Art
One possible prior art could be the use of GANs for data generation and augmentation in machine learning applications. Researchers have explored similar techniques for generating synthetic data in various domains.
Unanswered Questions
How does the performance of the combined GAN compare to individual GANs in terms of data generation quality?
The article does not provide information on the comparative performance of the combined GAN versus individual GANs in terms of data generation quality.
What are the specific user-defined constraints that can be imposed on the fabricated structured data?
The article does not detail the specific user-defined constraints that can be imposed on the fabricated structured data.
Original Abstract Submitted
A computer-implemented method including: training a first Generative Adversarial Network (GAN) based on original structured data; training a second GAN based on fabricated structured data that adhere to user-defined constraints; combining the first and second GANs into a combined GAN; training the combined GAN; and operating the trained combined GAN to generate new fabricated data that both imitate characteristics of the original structured data, and adhere to the user-defined constraints.
