Imaging Science Seminar: Recent Advances in Dynamic Tensor Analysis Based on Absolute Projections

Event Image
imaging science seminar cis panos

Recent Advances in Dynamic Tensor Analysis Based on Absolute Projections

Dr. Panos P. Markopoulos
Assistant Professor of Electrical Engineering
Kate Gleason College of Engineering, RIT

Register Here for Zoom Link

Abstract:
Data collections across diverse sensing modalities or configurations are naturally stored and processed in the form of N-way arrays, also known as tensors. Tucker decomposition is a standard multi-way generalization of Principal-Component Analysis (PCA), appropriately extended for the component analysis tensor data. Similar to PCA, Tucker decomposition has been shown to be sensitive against faulty data, due to its L2-norm-based formulation which places squared emphasis to peripheral/outlying entries. In this talk, we we review theory and efficient/dynamic algorithms for L1-Tucker, an L1-norm based reformulation of Tucker decomposition. Our experimental studies show that L1-Tucker attains similar performance to standard Tucker when the processed data tensor is corruption free, while it exhibits sturdy resistance against heavily corrupted entries. Thus, L1-Tucker could replace Tucker in an array of applications that rely on the analysis of possibly corrupted data.

Speaker Bio:
Dr. Panos P. Markopoulos is an Assistant Professor of Electrical Engineering at the Rochester Institute of Technology (RIT), Rochester NY, where he directs the Machine Learning Optimization and Signal Processing Laboratory (MILOS LAB - http://miloslab.org). In the summers of 2018 and 2020, he was a Visiting Research Faculty at the U.S. Air Force Research Laboratory (AFRL), in Rome NY. He received the Ph.D. degree in Electrical Engineering from The State University of New York at Buffalo, Buffalo NY, USA, in 2015. Dr. Markopoulos's research is in the areas of statistical signal processing, optimization, and machine learning, with a current focus on tensors, robustness, Lp-norm formulations, and dynamic learning. In these areas, he has co-authored multiple journal and conference articles. Dr. Markopoulos’s research has been supported with multiple grants from the U.S. National Science Foundation, the U.S. Government, the U.S. Air Force Office of Scientific Research, U.S. Air Force Research Lab, as well as industry partners. In 2020, he received the prestigious AFOSR Young Investigator Award. He is a member of IEEE Signal Processing, Computer, and Communications Societies, with high service activity that includes the organization of multiple conference events, including the 2019 IEEE International Workshop on Machine Learning for Signal Processing, the 2019-2021 versions of the SPIE DCS Conference on Big Data: Learning Analytics and Applications, and the 2017-2020 versions of the IEEE International Workshop on Wireless Communications and Networking in Extreme Environments.

Intended Audience:
Beginners, undergraduates, graduates, experts. Those with interest in the topic.


Contact
Marci Sanders-Arnett
Event Snapshot
When and Where
November 04, 2020
3:30 pm - 4:30 pm
Room/Location: See Zoom Registration Link
Who

This is an RIT Only Event

Interpreter Requested?

No

Topics
faculty
imaging science
interdisciplinary studies
rankings