20230106136. ELECTRONIC APPARATUS AND GENERATING METHOD OF TARGET DOMAIN simplified abstract (Samsung Electronics Co., Ltd.)
ELECTRONIC APPARATUS AND GENERATING METHOD OF TARGET DOMAIN
Organization Name
Inventor(s)
ELECTRONIC APPARATUS AND GENERATING METHOD OF TARGET DOMAIN - A simplified explanation of the abstract
This abstract first appeared for US patent application 20230106136 titled 'ELECTRONIC APPARATUS AND GENERATING METHOD OF TARGET DOMAIN
Simplified Explanation
The abstract describes an electronic apparatus that generates a target domain using a generative adversarial network (GAN) and a method for generating this target domain. The method involves reconstructing source data from a source domain, training the reconstructed source data to generate target data based on the source data, and creating a target domain that includes the generated target data. During the training process, the method identifies a loss value between a class loss value and a distance loss value, and applies this loss value to the reconstructed source data.
- The electronic apparatus generates a target domain using a generative adversarial network (GAN).
- The method involves reconstructing source data from a source domain.
- The reconstructed source data is trained to generate target data based on the source data.
- The generated target data is used to create a target domain.
- During training, the method identifies a loss value between a class loss value and a distance loss value.
- This loss value is applied to the reconstructed source data.
Potential Applications
- Image generation and manipulation
- Data augmentation for machine learning models
- Synthetic data generation for training AI algorithms
Problems Solved
- Lack of diverse and representative data for training machine learning models
- Difficulty in generating realistic and high-quality synthetic data
- Limited availability of labeled data for specific domains
Benefits
- Enables the generation of realistic and diverse data in a target domain
- Improves the performance and generalization of machine learning models
- Reduces the reliance on large amounts of labeled data for training AI algorithms
Original Abstract Submitted
an electronic apparatus generating a generative adversarial network (gan)-based target domain and a generating method thereof is provided. the generating method includes reconstructing source data included in a source domain, generating target data by training the reconstructed source data based on the source data, and generating a target domain including the generated target data, and the training comprises identifying at least one loss value between a class loss value by class loss and a distance loss value by distance matrix loss and applying at least one loss value between the identified class loss value and the distance loss value to the reconstructed source data.
