Samsung display co., ltd. (20240242494). FUSION MODEL TRAINING USING DISTANCE METRICS simplified abstract
Contents
FUSION MODEL TRAINING USING DISTANCE METRICS
Organization Name
Inventor(s)
Janghwan Lee of Pleasanton CA (US)
Jinghua Yao of San Jose CA (US)
Sai Marapareddy of Newark CA (US)
FUSION MODEL TRAINING USING DISTANCE METRICS - A simplified explanation of the abstract
This abstract first appeared for US patent application 20240242494 titled 'FUSION MODEL TRAINING USING DISTANCE METRICS
The abstract presents a method and system for controlling the performance of a fusion model by combining two neural networks.
- Obtaining a first set and a second set of candidate models pre-trained with different sources for the first and second neural networks.
- Determining a model distance for each possible pairing of candidate models from the first and second neural networks.
- Selecting a subset of pairings based on the model distance between them.
- Combining the selected pairings to generate two branches for a fusion model neural network.
- Potential Applications:**
- This technology can be applied in various fields such as image recognition, natural language processing, and predictive analytics. - It can enhance the performance of machine learning models in complex tasks requiring multiple neural networks.
- Problems Solved:**
- Improves the efficiency and accuracy of fusion models by selecting optimal pairings of candidate models. - Streamlines the process of combining neural networks for enhanced performance.
- Benefits:**
- Increased accuracy and reliability of fusion models. - Enhanced performance in tasks requiring the collaboration of multiple neural networks.
- Commercial Applications:**
- This technology can be utilized in industries such as healthcare, finance, and autonomous vehicles for advanced data analysis and decision-making processes.
- Prior Art:**
- Researchers have explored various methods for combining neural networks, but this specific approach of selecting candidate models based on model distance is a novel innovation.
- Frequently Updated Research:**
- Stay updated on advancements in neural network fusion techniques and applications in diverse industries to leverage the full potential of this technology.
- Questions about Neural Network Fusion:**
1. How does the selection of candidate models based on model distance impact the performance of fusion models? 2. What are the key considerations when combining neural networks for different applications?
Original Abstract Submitted
a method and a system are presented for controlling a performance of a fusion model. the method includes obtaining a first set and a second set of candidate models for a first and second neural networks, respectively. each of the first and second set of candidate models is pre-trained with a first source and a second source, respectively. for each possible pairing of one candidate model from the first neural network and one candidate model from the second neural network, a model distance dis determined. a subset of possible pairings of one first candidate model and one second candidate model is selected based on the model distance dbetween them. using the subset of possible parings, the first neural network and the second neural network are combined to generate two branches for a fusion model neural network.
