20230125839. METHOD AND APPARATUS FOR GENERATING SYNTHETIC DATA simplified abstract (SAMSUNG SDS CO., LTD.)
METHOD AND APPARATUS FOR GENERATING SYNTHETIC DATA
Organization Name
Inventor(s)
Jaehoon Lee of Gyeongsangnam-do (KR)
METHOD AND APPARATUS FOR GENERATING SYNTHETIC DATA - A simplified explanation of the abstract
This abstract first appeared for US patent application 20230125839 titled 'METHOD AND APPARATUS FOR GENERATING SYNTHETIC DATA
Simplified Explanation
The abstract describes a patent application for a synthetic data generating apparatus based on a generative adversarial network (GAN). The apparatus includes a generating unit that uses an invertible neural network to generate a fake data embedding vector from an original data embedding vector. A discriminating unit then determines whether the original data embedding vector and the fake data embedding vector are fake data.
- The apparatus uses a generative adversarial network (GAN) to generate synthetic data.
- It includes a generating unit that uses an invertible neural network to create fake data from original data.
- A discriminating unit is used to determine whether the generated data is fake or not.
Potential Applications
- Data augmentation for machine learning models.
- Privacy-preserving data sharing.
- Synthetic data generation for testing and validation purposes.
Problems Solved
- Lack of sufficient training data for machine learning models.
- Privacy concerns when sharing sensitive data.
- Difficulty in obtaining real-world data for testing and validation.
Benefits
- Increased availability of training data for machine learning models.
- Enhanced privacy protection by using synthetic data instead of real data.
- Cost-effective and efficient testing and validation using synthetic data.
Original Abstract Submitted
a generative adversarial network (gan)-based synthetic data generating apparatus according to an embodiment may include a generating unit configured to receive an original data embedding vector and to generate a fake data embedding vector by using an invertible neural network, and a discriminating unit configured to receive the original data embedding vector and the fake data embedding vector and to discriminate whether the original data embedding vector and the fake data embedding vector are fake data.