20230186098. DISTRIBUTED GENERATIVE ADVERSARIAL NETWORKS SUITABLE FOR PRIVACY-RESTRICTED DATA simplified abstract (Rutgers, The State University of New Jersey)
DISTRIBUTED GENERATIVE ADVERSARIAL NETWORKS SUITABLE FOR PRIVACY-RESTRICTED DATA
Organization Name
Rutgers, The State University of New Jersey
Inventor(s)
Dimitris Metaxas of Princeton NJ (US)
DISTRIBUTED GENERATIVE ADVERSARIAL NETWORKS SUITABLE FOR PRIVACY-RESTRICTED DATA - A simplified explanation of the abstract
This abstract first appeared for US patent application 20230186098 titled 'DISTRIBUTED GENERATIVE ADVERSARIAL NETWORKS SUITABLE FOR PRIVACY-RESTRICTED DATA
Simplified Explanation
The abstract describes an invention called Asynchronous Distributed Generative Adversarial Network (AsyndGAN) that includes a central computing system and multiple discriminator nodes. Here is a simplified explanation of the abstract:
- The AsyndGAN consists of a central computing system and at least two discriminator nodes.
- The central computing system has a generator neural network, an aggregator, and a network interface.
- Each discriminator node has its own training data set, and different nodes can use different types of data.
- The central computing system communicates with the discriminator nodes through the network interface.
- The central computing system aggregates data received from the discriminator nodes using the aggregator.
- The aggregated data is used to update the model for the generator neural network during training.
- The central computing system also includes a data access system that allows third parties to access synthetic data generated by the generator neural network.
Potential applications of this technology:
- Synthetic data generation for training machine learning models.
- Improving the performance of generative models by using multiple discriminator nodes.
- Enabling the use of different data modalities in the training process.
Problems solved by this technology:
- Improves the training process of generative models by using an asynchronous distributed approach.
- Allows for the use of different types of data in the training process.
- Provides a centralized system for aggregating data from multiple discriminator nodes.
Benefits of this technology:
- Enhances the quality and diversity of synthetic data generated by the generator neural network.
- Enables faster and more efficient training of generative models.
- Facilitates collaboration and third-party access to synthetic data.
Original Abstract Submitted
an asynchronous distributed generative adversarial network (asyndgan) can include a central computing system and at least two discriminator nodes. the central computing system can include a generator neural network, an aggregator, and a network interface. each discriminator node can have its own corresponding training data set. in addition, different discriminator nodes can use different data modalities. the central computing system communicates with each of the at least two discriminator nodes via the network interface and aggregates data received from the at least two discriminator nodes, via the aggregator, to update a model for the generator neural network during training of the generator neural network. the central computing system can further include a data access system that supports third party access to synthetic data generated by the generator neural network.
