University of Minnesota
University Relations
http://www.umn.edu/urelate
612-624-6868
myU OneStop


Go to unit's home.

Home | Seminars and Symposia | Upcoming seminars/symposia: Tuesday, October 17, 2017

An Introduction to Convolutional Neural Networks

by

Panagiotis (Panos) Stanitsas
University of Minnesota

Tuesday, October 17, 2017
11:00 a.m. lunch (provided) and social
12:00 p.m. seminar

401/402 Walter Library

add to google calendar

Seminar on Advances in Autonomy and Signal and Image Processing

Sponsored by: The Twin Cities Chapter of IEEE Aerospace and Electronic Systems Society and the University of Minnesota

The first part of this talk is concerned with exposing in some more detail fundamental components of Convolutional Neural Networks (CNNs) as well as popular ways to train a network. The neural network related literature that appeared in the 1990’s will be contrasted against CNNs in their present form towards highlighting research efforts that made this transition possible, shaping a major part of the area of deep learning. The second part of this talk discusses recent findings at the intersection of active and deep learning towards creating annotation-load aware schemes for training CNNs. CNNs typically require large amounts of annotated data to be trained effectively. However, in several scientific disciplines, including medical image analysis, generating such large annotated datasets requires specialized domain knowledge, and hence is usually very expensive. A novel application of active learning to data sample selection for training CNNs for cancerous tissue recognition is discussed in further detail.

 

Panagiotis Stanitsas received his Diploma of Engineering in Civil Engineering from the University of Patras, Greece in 2010. He received his M.S. in Project Management and Transportation from the University of Patras, Greece in 2011 and his M.Sc. in Civil Engineering from the University of Minnesota in 2013. Currently, he is a Ph.D. candidate at the department of Computer Science and Engineering at the University of Minnesota advised by Professor Nikolaos Papanikolopoulos and Dr. Vassilios Morellas. His research focuses on the development of machine learning methods with emphasis on cancerous tissue recognition. He was awarded with an award for excellence in Transportation from the University of Patras in 2010, while he has also received the interdisciplinary Matthew J. Huber award of excellence at the University of Minnesota in 2013.