Imaging Science Seminar: Dr. Bartosz Krawczyk

Imaging Science Seminar
Learning from imbalanced and streaming data for computer vision
Dr. Bartosz Krawczyk
Assistant Professor, Department of Computer Science
Virginia Commonwealth University

This talk will give an overview of the imbalanced learning domain, focusing on contemporary challenging scenarios and recent developments. Special attention will be given to data-level difficulties and understanding minority classes, multi-class imbalanced problems, and continual learning from data streams with dynamically evolving classes. The talk will discuss various resampling methods, low-dimensional embeddings, ensemble learning approaches, and Explainable Artificial Intelligence tools that I developed to efficiently handle such challenging scenarios.

Abstract: Learning from imbalanced data is considered one of the vital challenges in contemporary machine learning and computer vision. Underlying bias in data and skewed distributions have severely negative impact on classifiers. Imbalance problem was first recognized almost three decades ago and has remained a critical challenge. The advent of deep learning has led to emergence of novel challenges, as these models memorize training data, which hurts their ability to generalize to under-represented classes. This talk will give an overview of the imbalanced learning domain, focusing on contemporary challenging scenarios and recent developments. Special attention will be given to data-level difficulties and understanding minority classes, multi-class imbalanced problems, and continual learning from data streams with dynamically evolving classes. The talk will discuss various resampling methods, low-dimensional embeddings, ensemble learning approaches, and Explainable Artificial Intelligence tools that I developed to efficiently handle such challenging scenarios. I will position this research in the context of imaging science, highlighting this field’s need for robust, skew-insensitive, and streaming machine learning.

Speaker Bio: Bartosz Krawczyk is an Assistant Professor in the Department of Computer Science, Virginia Commonwealth University, Richmond VA, where he heads the Machine Learning and Stream Mining Lab. He obtained his M.Sc. and Ph.D. degrees from Wroclaw University of Science and Technology, Poland, in 2012 and 2015 respectively. Dr. Krawczyk’s current research interests include machine learning, data streams, continual learning, class imbalance, and explainable artificial intelligence. He has authored more than 60 journal papers and over 100 contributions to conferences, as well as co-authored the book “Learning from Imbalanced Data Sets” (Springer, 2018). Dr. Krawczyk’s research has been funded by industrial grants from Amazon and Ho-Ho-Kus Inc. He was a recipient of several awards for his scientific achievements, such as IEEE Richard Merwin Scholarship, IEEE Outstanding Leadership Award, and Amazon AWS Machine Learning Award, among others. Dr. Krawczyk was an invited keynote speaker at six conferences and workshops. He served as a Guest Editor for four journal special issues, as a Chair for fifteen special session and workshops, and gave tutorials at IEEE DSAA 2019 and IEEE BigData 2020. Dr. Krawczyk is a member of the Program Committee for conferences, such as KDD (Senior PC), AAAI, IJCAI, ECML-PKDD, PAKDD, IEEE BigData, and IJCNN. He is a member of the editorial board for Applied Soft Computing (Elsevier).

Intended Audience: Undergraduates, graduates, and experts. Those with interest in the topic.