20240054347. SYSTEM AND METHOD OF TRAINING A NEURAL NETWORK MODEL simplified abstract (TECHNION RESEARCH & DEVELOPMENT FOUNDATION LIMITED)
Contents
SYSTEM AND METHOD OF TRAINING A NEURAL NETWORK MODEL
Organization Name
TECHNION RESEARCH & DEVELOPMENT FOUNDATION LIMITED
Inventor(s)
SYSTEM AND METHOD OF TRAINING A NEURAL NETWORK MODEL - A simplified explanation of the abstract
This abstract first appeared for US patent application 20240054347 titled 'SYSTEM AND METHOD OF TRAINING A NEURAL NETWORK MODEL
Simplified Explanation
The patent application describes a method and system for implementing a machine-learning based function using a neural network model.
- Providing a neural network model with multiple parameters.
- Training the neural network model over multiple epochs using a training dataset to implement a predefined machine learning function.
- Calculating a profile vector for each neural network parameter to represent its evolution through the training epochs.
- Approximating the value of each neural network parameter based on the profile vector.
- Replacing the original parameter values in the trained neural network model with the calculated approximated values to obtain an approximated version of the model.
---
- Potential Applications
- Improving the efficiency and accuracy of machine learning models.
- Enhancing the performance of neural networks in various applications such as image recognition, natural language processing, and predictive analytics.
- Problems Solved
- Overfitting of neural network models.
- Reducing the computational resources required for training neural networks.
- Improving the interpretability of neural network models.
- Benefits
- Increased accuracy and reliability of machine learning models.
- Faster training times for neural networks.
- Better understanding of the behavior of neural network parameters.
Original Abstract Submitted
a method and system for implementing a machine-learning (ml) based function may include providing a nn model comprising a plurality of nn parameters; training the nn model over a plurality of training epochs, to implement a predefined ml function, based on a training dataset; for one or more nn parameters of the plurality of nn parameters: (i) calculating a profile vector, representing evolution of the nn parameter through the plurality of training epochs; and (ii) calculating an approximated value of the at least one nn parameter, based on the profile vector; and replacing at least one nn parameter value in the trained nn model with a respective calculated approximated value, to obtain an approximated version of the trained nn model.