18447836. METHOD AND APPARATUS FOR SIMULTANEOUS TRAINING AND CORRECTION OF ARTIFICIAL NEURAL NETWORK AND DATASET simplified abstract (ELECTRONICS AND TELECOMMUNICATIONS RESEARCH INSTITUTE)
METHOD AND APPARATUS FOR SIMULTANEOUS TRAINING AND CORRECTION OF ARTIFICIAL NEURAL NETWORK AND DATASET
Organization Name
ELECTRONICS AND TELECOMMUNICATIONS RESEARCH INSTITUTE
Inventor(s)
METHOD AND APPARATUS FOR SIMULTANEOUS TRAINING AND CORRECTION OF ARTIFICIAL NEURAL NETWORK AND DATASET - A simplified explanation of the abstract
This abstract first appeared for US patent application 18447836 titled 'METHOD AND APPARATUS FOR SIMULTANEOUS TRAINING AND CORRECTION OF ARTIFICIAL NEURAL NETWORK AND DATASET
Simplified Explanation:
This patent application describes a method for mapping data from different domains to a common joint embedding space using a neural network.
- Training a mapping neural network to create a joint embedding space.
- Generating a prediction matrix for an input dataset using the neural network.
- Creating a merging dictionary to correct the input dataset by merging classes from the prediction matrix.
Key Features and Innovation:
- Utilizes a neural network to map data from diverse domains to a shared embedding space.
- Improves the accuracy and efficiency of data mapping processes.
- Enables the correction of input datasets by merging classes based on prediction matrices.
Potential Applications:
This technology can be applied in fields such as:
- Machine learning
- Data analysis
- Information retrieval
Problems Solved:
- Simplifies the process of mapping data from different domains.
- Enhances the accuracy of data mapping and classification tasks.
Benefits:
- Improved data mapping accuracy.
- Enhanced data classification capabilities.
- Increased efficiency in data analysis tasks.
Commercial Applications:
Potential commercial applications include:
- Data analytics software
- Machine learning platforms
- Information retrieval systems
Questions about Mapping Data to a Joint Embedding Space:
1. How does the neural network improve the accuracy of data mapping? 2. What are the potential challenges in implementing this technology in real-world applications?
Original Abstract Submitted
An embodiment of the present disclosure discloses a method of operating a computing device for mapping data from different domains to a common joint embedding space, and the method of operating a computing device includes training a mapping neural network constituting a joint embedding space using an input dataset, generating a prediction matrix of an input dataset using the mapping neural network, and generating a merging dictionary merging classes from the prediction matrix to correct the input dataset.
