18276016. BUILDING AN EXPLAINABLE MACHINE LEARNING MODEL simplified abstract (Telefonaktiebolaget LM Ericsson (publ))

BUILDING AN EXPLAINABLE MACHINE LEARNING MODEL

Organization Name

Telefonaktiebolaget LM Ericsson (publ)

Inventor(s)

Perepu Satheesh Kumar of Chennai (IN)

M Saravanan of Chennai (IN)

Sai Hareesh Anamandra of BANGALORE KARNATAKA (IN)

BUILDING AN EXPLAINABLE MACHINE LEARNING MODEL - A simplified explanation of the abstract

This abstract first appeared for US patent application 18276016 titled 'BUILDING AN EXPLAINABLE MACHINE LEARNING MODEL

Simplified Explanation

The abstract describes a computer-implemented method for building a machine learning model, where the model includes multiple layers with filters, and class labels are used in training and determining filter values.

• Training a machine learning model with input data and class labels
• Obtaining output data with class probabilities
• Determining working values for filters in each layer using class labels and probabilities
• Identifying dominant filters in each layer based on a threshold
• Building a subset model based on dominant filters in each layer

Potential Applications

This technology could be applied in various fields such as image recognition, natural language processing, and predictive analytics.

Problems Solved

This method helps in improving the efficiency and accuracy of machine learning models by focusing on dominant filters in each layer, potentially reducing computational resources and enhancing performance.

Benefits

The benefits of this technology include faster model training, better interpretability of model decisions, and potentially higher accuracy in classification tasks.

Potential Commercial Applications

Potential commercial applications of this technology could include developing advanced recommendation systems, fraud detection algorithms, and personalized marketing strategies.

Possible Prior Art

One possible prior art could be the use of filter importance analysis in deep learning models to improve model performance and interpretability.

How does this method compare to existing techniques for filter selection in machine learning models?

This article does not provide a direct comparison with existing techniques for filter selection in machine learning models. It would be interesting to know if this method outperforms traditional approaches in terms of model performance and efficiency.

What are the potential limitations or challenges of implementing this method in real-world applications?

The article does not address potential limitations or challenges of implementing this method in real-world applications. It would be important to consider factors such as scalability, computational resources, and generalizability when applying this technique in practical settings.

Original Abstract Submitted

A computer-implemented method for building a machine learning (ML) model is provided. The method includes training a ML model using a set of input data, wherein the ML model includes a plurality of layers and each layer includes a plurality of filters, and wherein the set of input data includes class labels; obtaining a set of output data from training the ML model, wherein the set of output data includes class probabilities values; determining, for each layer in the ML model, by using the class labels and the class probabilities values, a working value for each filter in the layer; determining, for each layer in the ML model, a dominant filter, wherein the dominant filter is determined based on whether the working value for the filter exceeds a threshold; and building a subset ML model based on each dominant filter for each layer, wherein the subset ML model is a subset of the ML model.