Michalis Papakostas

Research Scientist

GN Advanced Science

R&D Global Research Group, GN Group


I am a Research Scientist working on HCI, Machine Learning and Human-Factors Engineering concepts at GN Group's, Advanced Science team with a focus on hearing-aid related technology.

Previously I was a Postdoctoral Fellow at the University of Michigan's AI Lab, doing research on Multi-Modal Processing for Human Behavior Modeling at the Language and Information Technology Group directed by Professor Rada Mihalcea.

I received my Ph.D. in 2019 as a member of the Heracleia- Human Centered Computing Lab in a joint Doctorate program between the University of Texas at Arlington and the National Center of Scientific Research, "Demokritos", under the supervision of Professor Fillia Makedon & Dr. Vangelis Karkaletsis . Before that, I received a M.Eng. Diploma from the Department of Electrical & Computer Engineering at the Technical University of Crete in 2013. My M.Eng. Thesis was supervised by Professor Vassilis Digalakis.

My primary research interests are on the fields of Multi-Modal Processing, Human Machine Interaction and Machine Learning with a special focus on User Behavior Understanding, Modeling & Monitoring.


Email: mpapakostas@gnhearing.com

News

  • April 2021 - Outstanding Paper Award and Best Paper Award Finalist @ IUI2021 for our paper "Understanding Driving Distractions: A Multimodal Approach to Distraction Detection and Recognition."

  • January 2021 - I will be joining GN Group's outstanding research team on developing novel HCI solutions for tomorrow's hearing-aid assistive technologies!

  • December 2020 - Paper accepted at IUI 2021 "Understanding Driving Distractions: A Multimodal Approach to Distraction Detection and Recognition.

  • November 2020 - I will be giving a talk at the University of Texas at Arlington on November 5th

  • August 2020 - I will be giving a talk at the ACM Chicago CHI Chapter on September 3rd

  • February 2020 - Our paper with title "Towards detecting levels of alertness in drivers using multiple modalities" was accepted for publication at PETRA 2020