18458709. MODEL COMPRESSION BY SPARSITY-INDUCING REGULARIZATION OPTIMIZATION simplified abstract (Microsoft Technology Licensing, LLC)
MODEL COMPRESSION BY SPARSITY-INDUCING REGULARIZATION OPTIMIZATION
Organization Name
Microsoft Technology Licensing, LLC
Inventor(s)
Tianyi Chen of Redmond WA (US)
MODEL COMPRESSION BY SPARSITY-INDUCING REGULARIZATION OPTIMIZATION - A simplified explanation of the abstract
This abstract first appeared for US patent application 18458709 titled 'MODEL COMPRESSION BY SPARSITY-INDUCING REGULARIZATION OPTIMIZATION
Simplified Explanation
The abstract describes a method for improving the performance of neural networks by compressing them through a sparsity-inducing regularization optimization process. The compressed machine learning model is trained using a set of training data and then executed to generate outputs.
- The method involves compressing neural networks to improve their performance.
- A sparsity-inducing regularization optimization process is used to generate a compressed machine learning model.
- The compressed model is trained using a set of training data.
- The compressed model is then executed to generate one or more outputs.
Potential Applications
- This technology can be applied in various fields where neural networks are used, such as image recognition, natural language processing, and recommendation systems.
- It can be used to improve the efficiency and speed of neural networks in real-time applications, such as autonomous vehicles and robotics.
Problems Solved
- The method addresses the problem of limited performance of neural networks due to the number of operations being performed and data management among different memory components.
- It solves the problem of optimizing neural networks to reduce their size and computational requirements without sacrificing accuracy.
Benefits
- The compressed machine learning models generated through this method can significantly reduce the computational resources required for training and execution.
- The improved efficiency and speed of the compressed models make them suitable for real-time applications.
- The method allows for better utilization of memory components in neural networks, leading to improved performance.
Original Abstract Submitted
The performance of a neural network (NN) and/or deep neural network (DNN) can limited by the number of operations being performed as well as management of data among the various memory components of the NN/DNN. A sparsity-inducing regularization optimization process is performed on a machine learning model to generate a compressed machine learning model. A machine learning model is trained using a first set of training data. A sparsity-inducing regularization optimization process is executed on the machine learning model. Based on the sparsity-inducing regularization optimization process, a compressed machine learning model is received. The compressed machine learning model is executed to generate one or more outputs.
