Intel corporation (20240119710). METHODS, SYSTEMS, APPARATUS, AND ARTICLES OF MANUFACTURE TO AUGMENT TRAINING DATA BASED ON SYNTHETIC IMAGES simplified abstract
Contents
- 1 METHODS, SYSTEMS, APPARATUS, AND ARTICLES OF MANUFACTURE TO AUGMENT TRAINING DATA BASED ON SYNTHETIC IMAGES
- 1.1 Organization Name
- 1.2 Inventor(s)
- 1.3 METHODS, SYSTEMS, APPARATUS, AND ARTICLES OF MANUFACTURE TO AUGMENT TRAINING DATA BASED ON SYNTHETIC IMAGES - 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 Original Abstract Submitted
METHODS, SYSTEMS, APPARATUS, AND ARTICLES OF MANUFACTURE TO AUGMENT TRAINING DATA BASED ON SYNTHETIC IMAGES
Organization Name
Inventor(s)
Anmol Bhasin of SAS Nagar (IN)
Shekar Ramachandran of Bengaluru (IN)
Rudra Nath Palit of Kolkata (IN)
Rupali Agrahari of Sultanpur (IN)
Sai Pramod Gadam of Bengaluru (IN)
METHODS, SYSTEMS, APPARATUS, AND ARTICLES OF MANUFACTURE TO AUGMENT TRAINING DATA BASED ON SYNTHETIC IMAGES - A simplified explanation of the abstract
This abstract first appeared for US patent application 20240119710 titled 'METHODS, SYSTEMS, APPARATUS, AND ARTICLES OF MANUFACTURE TO AUGMENT TRAINING DATA BASED ON SYNTHETIC IMAGES
Simplified Explanation
The patent application describes methods, systems, apparatus, and articles of manufacture for augmenting training data using synthetic images. Specifically, it involves using a generative adversarial network (GAN) to generate images representing different racial domains to enhance a training dataset.
- Programmable circuitry generates a latent representation of an image from one racial domain using the GAN.
- The circuitry then generates an image representing a different racial domain based on the latent representation.
- The training dataset is expanded by incorporating the generated images.
Potential Applications
This technology could be applied in various fields such as computer vision, artificial intelligence, and machine learning for improving the diversity and accuracy of training datasets.
Problems Solved
This innovation addresses the issue of limited diversity in training data, which can lead to biased or inaccurate AI models. By generating synthetic images representing different racial domains, the training dataset becomes more inclusive and representative.
Benefits
The benefits of this technology include enhanced model performance, reduced bias, improved generalization, and increased fairness in AI applications.
Potential Commercial Applications
Potential commercial applications of this technology include AI-powered systems for facial recognition, object detection, autonomous vehicles, and medical imaging, where diverse and representative training data is crucial for optimal performance.
Possible Prior Art
One possible prior art could be the use of data augmentation techniques in machine learning to increase the diversity of training datasets. Another could be the use of GANs for generating synthetic images to improve model performance.
Original Abstract Submitted
methods, systems, apparatus, and articles of manufacture to augment training data based on synthetic images are disclosed. an example apparatus disclosed herein includes programmable circuitry to generate, with one or more first layers of a generative adversarial network (gan), a latent representation corresponding to a first image representative of a first racial domain, generate, with one or more second layers of the gan, a second image based on the latent representation, the second image corresponding to a second racial domain different from the first racial domain, and augment a training dataset based on the second image.
