Presented By: Department of Computational Medicine and Bioinformatics DCMB
DCMB Tools and Technology Seminar by Matt Hodgman
“A novel, glass-box machine learning algorithm that leverages both knowledge and data to solve diverse and complex problems”
Abstract
Machine learning is becoming increasingly relevant and accessible in every field. But what does the machine actually learn? The demand for explainable machine learning has never been higher--especially in sensitive fields like healthcare. An explainable machine learning algorithm can describe the learned relationships between input data and output predictions, as well as describe the reasons for single specific prediction. The list of truly explainable machine learning algorithms is short and often limited by poor predictive performance at complex tasks. We have developed a novel fuzzy neural network that exhibits both high explainability and accuracy in a variety of applications from classifying disease outcomes to regressing drug doses. This algorithm can use both domain-expert knowledge and complex data to converge on complex solutions. We present how this algorithm works for new users, the depth of explainability it offers, and highlight its performance in various applications.
About the DCMB Tools & Technology Seminar Series
The DCMB Tools and Technology Seminar Series is held in Medical Science Building 1 (MS1), Room 4B700, each Thursday at 12pm EST. Each seminar highlights a computational tool, technology, or methodology that is under development or in current use and is of special interest to DCMB and University researchers. Presenters are U-M researchers and students.
These seminars are live-streamed and recorded and made available for future viewing via the DCMB YouTube Channel
Machine learning is becoming increasingly relevant and accessible in every field. But what does the machine actually learn? The demand for explainable machine learning has never been higher--especially in sensitive fields like healthcare. An explainable machine learning algorithm can describe the learned relationships between input data and output predictions, as well as describe the reasons for single specific prediction. The list of truly explainable machine learning algorithms is short and often limited by poor predictive performance at complex tasks. We have developed a novel fuzzy neural network that exhibits both high explainability and accuracy in a variety of applications from classifying disease outcomes to regressing drug doses. This algorithm can use both domain-expert knowledge and complex data to converge on complex solutions. We present how this algorithm works for new users, the depth of explainability it offers, and highlight its performance in various applications.
About the DCMB Tools & Technology Seminar Series
The DCMB Tools and Technology Seminar Series is held in Medical Science Building 1 (MS1), Room 4B700, each Thursday at 12pm EST. Each seminar highlights a computational tool, technology, or methodology that is under development or in current use and is of special interest to DCMB and University researchers. Presenters are U-M researchers and students.
These seminars are live-streamed and recorded and made available for future viewing via the DCMB YouTube Channel
Related Links
Explore Similar Events
-
Loading Similar Events...
