20230146468. SYSTEMS AND METHODS FOR A LIGHTWEIGHT PATTERN-AWARE GENERATIVE ADVERSARIAL NETWORK simplified abstract (Ceremorphic, Inc.)

SYSTEMS AND METHODS FOR A LIGHTWEIGHT PATTERN-AWARE GENERATIVE ADVERSARIAL NETWORK

Organization Name

Ceremorphic, Inc.

Inventor(s)

Chandrajit Pal of Khandi (IN)

Manmohan Tripathi of Hyderabad (IN)

Govardhan Mattela of Hyderabad (IN)

SYSTEMS AND METHODS FOR A LIGHTWEIGHT PATTERN-AWARE GENERATIVE ADVERSARIAL NETWORK - A simplified explanation of the abstract

This abstract first appeared for US patent application 20230146468 titled 'SYSTEMS AND METHODS FOR A LIGHTWEIGHT PATTERN-AWARE GENERATIVE ADVERSARIAL NETWORK

Simplified Explanation

The abstract describes a computer-implemented method involving the training of a generative adversarial network (GAN) on one or more processors. The method includes several steps:

• Pattern extraction is applied to a set of training data to extract feature embeddings representing features of the data.
• The feature embeddings are attenuated to create attenuated feature embeddings.
• The attenuated feature embeddings are provided to a generator of the GAN to partly control the generation of synthetic data.
• The generator uses the attenuated embeddings to generate synthetic data.

Potential applications of this technology:

• Synthetic data generation for various purposes such as data augmentation, privacy protection, or simulation.
• Training machine learning models on larger and more diverse datasets by generating synthetic data.

Problems solved by this technology:

• Limited availability of real-world data for training machine learning models.
• Privacy concerns when using sensitive or personal data.
• Lack of diversity in training datasets.

Benefits of this technology:

• Enables the generation of synthetic data that closely resembles real-world data.
• Provides a solution for data scarcity and privacy concerns.
• Enhances the performance and robustness of machine learning models by training on more diverse datasets.

Original Abstract Submitted

a computer-implemented method includes training at least a generative adversarial network, the method operable on one or more processors. the method includes at least (1) applying pattern extraction to a set of training data to extract one or more feature embeddings representing one or more features of the training data, (2) attenuating the one or more feature embeddings to create one or more attenuated feature embeddings, (3) providing the one or more attenuated embeddings to a generator of the generative adversarial network as a condition to at least partly control the generator in generating synthetic data, the providing being performed automatically and dynamically during training of the generator, and (4) with the generator, generating synthetic data based at least in part on the attenuated embeddings.