Soheil Feizi Receives the AWS Machine Learning Research Award

Assistant Professor Soheil Feizi recently won the AWS Machine Learning Research Award for his project titled “Explainable Deep Learning: Accuracy, Robustness and Fairness”.

 The AWS Machine Learning Research Awards (MLRA) aims to support research to advance the frontiers of machine learning (ML) and its application across a wide range of problems.

The funded project focuses on developing accurate, robust and fair interpretation methods for deep learning models.

“The growing use of deep learning in sensitive applications such as medicine, autonomous driving, and finance raises concerns about human trust in machine learning systems”, said Feizi. For trained models, a central question is test-time interpretability: how humans can understand the reasoning behind model predictions, he explains.

Feizi’s research project aims at identifying the importance of each input feature for a model’s prediction.

In addition to the unrestricted cash award and  access to cutting-edge infrastructure and managed services through AWS promotional credits the award offers opportunities to participate in AWS events and receive live one-on-one training sessions with AWS data scientists and engineers.

In addition to being an Assistant Professor of Computer Science, Feizi holds joint appointments in UMIACS and University of Maryland Center for Machine Learning.