Dr. Angshul Majumdar of Indraprastha Institute of Information Technology, Delhi

Deep learning has recently attracted a lot of attention from academia and industry. In this talk we discuss the basic building blocks of deep learning – autoencoders, restricted Botlzmann machine and convolutional neural network. These three techniques are born out of the machine learning community. We also discuss about two powerful representation learning tools from the signal processing community – dictionary learning and transform learning. We show how in the recent past (last one year) dictionary learning has been used for building deeper architectures. We conclude the talk with the deeper versions of transform learning – a topic that is going to be realized in the near future.

Biography:

Angshul Majumdar received the Bachelor's degree from the Bengal Engineering College, Shibpur, India, and the Master's and Ph.D. degrees from the University of British Columbia in 2009 and 2012, respectively. He is currently an Assistant Professor with the Indraprastha Institute of Information Technology (IIIT), Delhi, India. He has co-authored over 150 papers in journals and reputed conferences. He has authored the book Compressed Sensing for Magnetic Resonance Image Reconstruction (Cambridge University Press, 2015) and co-edited a volume on MRI: Physics, Reconstruction, and Analysis (CRC Press, 2015). His research interests are broadly in the areas of signal processing and machine learning, with a specific focus on deep learning. He is a Senior Member of IEEE and is currently serving as the Chair of the IEEE SPS Chapter's Committee and the Chair of the IEEE SPS Delhi Chapter.

Email:

Address:Electronics and Communications Engineering, Indraprastha Institute of Information Technology, Delhi, Delhi, India, 110020

Dr. Angshul Majumdar of Indraprastha Institute of Information Technology, Delhi

Biography:

Email:

Address:Delhi, Delhi, India