17875937. 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 17875937 titled 'ELECTRONIC APPARATUS AND GENERATING METHOD OF TARGET DOMAIN
Simplified Explanation
The abstract describes an electronic apparatus and method for generating a target domain using a generative adversarial network (GAN). The method involves reconstructing source data from a source domain, training the reconstructed data to generate target data, and creating a target domain with the generated data. The training process includes identifying loss values related to class and distance, and applying these values to the reconstructed source data.
- The patent application describes an electronic apparatus and method for generating a target domain using a GAN.
- The method involves reconstructing source data, training the reconstructed data, and generating target data.
- The training process includes identifying loss values related to class and distance, and applying them to the reconstructed source data.
Potential Applications
This technology has potential applications in various fields, including:
- Image and video processing: The generated target domain can be used for image and video enhancement, synthesis, or manipulation.
- Data augmentation: The generated target domain can be used to increase the diversity and size of training datasets for machine learning models.
- Anomaly detection: The generated target domain can be used to detect anomalies or outliers in datasets.
Problems Solved
This technology addresses the following problems:
- Limited availability of target domain data: By generating a target domain using the GAN-based method, it overcomes the need for a large amount of labeled target data.
- Lack of diversity in training datasets: The generated target domain can provide additional diversity to training datasets, improving the performance of machine learning models.
- Difficulty in detecting anomalies: The generated target domain can help in identifying anomalies or outliers in datasets, aiding in anomaly detection tasks.
Benefits
The use of this technology offers several benefits:
- Improved data quality: The generated target domain can enhance the quality of data by synthesizing new samples that are similar to the source data.
- Increased training dataset size: By generating additional target data, the size of training datasets can be expanded, leading to better model performance.
- Enhanced anomaly detection: The generated target domain can improve the accuracy of anomaly detection algorithms by providing a more diverse and representative dataset.
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.
