Researchers develop explainability techniques to understand deep learning algorithms

Sherin Mathews, a Senior Data Scientist based in Santa Clara, California, is a revolutionary pioneer of the practical application of machine-learning. She recently presented at the Computing Conference London her latest work, “Explainable Artificial Intelligence Applications in NLP, Biomedical, and Malware Classification: A Literature Review.”

The paper is just the most recent in a career that explores the widespread applications of deep learning algorithms for many applications, which include biomedical ECG classification, malware detection, and health activity monitoring.

“Deep learning algorithms have achieved high-performance accuracy in complex domains such as image classification, face recognition sentiment analysis, text classification, and speech understanding,” the paper claims. “Due to the nested non-linear structure of deep learning algorithms, these highly successful models are usually applied in a black-box manner, i.e., no information is provided about what exactly causes them to arrive at their predictions. The effectiveness of these systems is thus limited by the machine’s current inability to explain its decisions and actions to human users.”

That lack of transparency is a major drawback of the technology, and Mathews looks to find out how to better utilize the algorithms through this unique approach.

The full abstract can be found at

With her work, Mathews has developed novel dictionary and deep learning algorithms for biomedical application for ECG classification and has also been instrumental in developing explainability frameworks to make machine learning algorithms more understandable.

This revolutionary work is motivated due to lacking transparency of the deep learning technology. She developed this unique technique on biomedical, text classification and malware detection domain. Her work is so adaptable that it can be extended and applied to all domains that make use of machine learning be it face recognition, text classification as demonstrated in her paper.

Mathews’ revolutionary research has been cited countless times and includes a variety of projects.

“A novel application of deep learning for single-lead ECG classification,” “Leveraging a discriminative dictionary learning algorithm for single-lead ECG classification,” “Maximum correntropy based dictionary learning framework for physical activity recognition using wearable sensors” and “Am I your sibling?’ Inferring kinship cues from facial image pairs,” are just a few of her most cited papers that are contributing to the development of practical applications for machine-based learning.

A more complete listing of Mathews’ work can be found on Google Scholar by clicking here

Media Contact
Contact Person: Sherin Mathews
Email: [email protected]
City: Santa Clara
State: CA 95051
Country: United States