The Journey from Shallow to Deep LearningDr. Anush Sankaran (IBM Research AI),   Dr. Chandra Mouli P V S (VIT),   Dr. Aditya Nigam (IIT Mandi)
The domain of computer vision and pattern recognition has been actively transformed by deep learning. In order to educate the audience on this growing need, this tutorial will walk the audience through a range of topics from the need and introduction of deep learning to the state-of-art technology, with plenty of hands-on assignments. One of the key highlights of this tutorial would be to demonstrate IBM’s DARVIZ, a free visual IDE to implement deep learning models within few minutes, without the need for writing a single line of code.
Computer Vision for Autonomous NavigationProf. C.V. Jawahar et. al. (IIIT Hyderabad)
We will start with the fundamentals of modern computer vision (primarily based on deep learning) with focus on classification, detection and segmentation. We will then discuss their applications in autonomous navigation and the special challenges and requirements in this domain. We will then move on to the problems of specific interest in autonomous navigation, and how computer vision supplements and complements other sensory inputs (eg. 3D). We will also discuss the emerging problems in autonomous navigation and special challenges in the Indian environments.
Deep Learning for Medical Image AnalysisDr. Debdoot Sheet (IIT Kharagpur)
This tutorial will focus on understanding the inherent hierarchy in solving most of the medical imaging and image analysis problems, then move over to the buzz surrounding this topic of deep learning and how firm does the buzz hold on to the claims it boasts of? Also we would host a hands-on tutorial with implementing a deep network using auto encoders for retinal vessel detection problems on the DRIVE and STARE datasets, and subsequently move on to using deep networks for feature discovery, using deep-hybrid architectures by coupling them with some ensemble learners like random forests, and finally end with how to address the dilemma of interpreting what the deep network has learnt and how it has learnt to do so? Following this we would end on a note to the major challenges in the field of deep learning existing today and some thought provoking research problems.
Organisation of the Tutorial: Concept of hierarchical learning in medical images (30 mins), Introduction to deep learning and common notions (30 mins), Autoencoders for representation learning in medical images+Lab (1 hour), Convolutional neural networks for medical image classification+Lab (1 hour), CNN for medical image segmentation and restoration+Lab (1 hour).
Lab sessions would be conducted on a remote HPC provided by Intel. Participants need to bring in their own laptops, wifi connectivity to be provided at the venue. If participants prefer to work out examples on their laptops, then they need to bring ones preloaded with Ubuntu Linux 16.04 OS, Anaconda Python 2.7 and Pytorch for Python2.7.