18623923. FUSION MODEL TRAINING USING DISTANCE METRICS simplified abstract (Samsung Display Co., LTD.)
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 18623923 titled 'FUSION MODEL TRAINING USING DISTANCE METRICS
The abstract presents a method and system for controlling the performance of a fusion model by combining pre-trained candidate models from two neural networks.
- Obtaining a first set and a second set of candidate models for two neural networks.
- Pre-training each candidate model with a specific data source.
- Determining model distances between pairs of candidate models.
- Selecting a subset of candidate model pairings based on model distances.
- Combining selected candidate models to create branches for a fusion model neural network.
Potential Applications: - Enhancing the performance of fusion models in various industries such as healthcare, finance, and autonomous vehicles. - Improving the accuracy and efficiency of data analysis and decision-making processes.
Problems Solved: - Optimizing the fusion model performance by selecting the best candidate models from multiple neural networks. - Streamlining the process of combining neural networks for improved results.
Benefits: - Increased accuracy and reliability of fusion models. - Enhanced decision-making capabilities in complex data analysis tasks. - Improved efficiency and effectiveness of neural network applications.
Commercial Applications: Title: "Enhancing Fusion Model Performance for Data Analysis Applications" This technology can be utilized in industries such as healthcare for medical diagnosis, finance for predictive analytics, and autonomous vehicles for real-time decision-making.
Prior Art: Readers can explore prior research on fusion models, neural network combinations, and model optimization techniques in the field of machine learning and artificial intelligence.
Frequently Updated Research: Stay updated on the latest advancements in fusion model optimization, neural network integration, and performance evaluation methods to enhance the application of this technology.
Questions about Fusion Model Performance: 1. How does the method of combining pre-trained candidate models improve the performance of fusion models? 2. What are the key factors to consider when selecting candidate model pairings for a fusion model?
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.
