This curriculum is developed to help participants appreciate the application of deep learning in the field of Computer Vision. The main objective of this course is to introduce Deep Learning and its relevance in the area of Computer Vision, while reviewing basic image processing techniques using various Python libraries.

This course provides an understanding on Neural Networks and Convolution Neural Networks. It focuses on selective areas of application of Deep Learning in Computer Vision such as Image Classification, Object Detection, Object Recognition, Content Based Image Retrieval, Image Generation, Action Recognition.

On completion of this course, participants will be able to recognize an image as a special form of data and appreciate the associated data handing techniques, describe a process for solving computer vision problems, classify computer vision problems into standard typology, develop codes in Python and PyTorch (deep learning architecture) for Computer Vision applications, correlate and assess the solution approach followed.

In this course participants will be introduced to the burgeoning world of Blockchain technology. The fundamental elements of blockchain are explained in detail. Key topics covered during the program include cryptography, hash functions, consensus, proof of stake and Merkle trees.

Blockchain, in layman terms, is a shared digital ledger for recording transactions and tracking assets. This course opens the door to developing democratic systems with the distributed consensus in the online digital world today. Learners will appreciate how blockchain paves way for several applications, including logging transactions in the ledger in a decentralized manner. This course also reveals the current trends in financial services, health care, enterprise resource planning, land record management, digital certificate management etc. through the concepts of blockchain.

Optimization is at the core of all machine learning techniques. In this course, you will be introduced to the fundamental concepts of optimization. The core idea of optimization is to find solutions that help maximize or minimize the objective of interest through a correct choice for parameter values.Participants will be taught constrained and unconstrained optimization, gradient descent, and variants such as stochastic gradient descent, linear and mixed-integer programming. 

Optimization techniques have been widely used in different domains to resolve complex problems like scheduling and resource allocation. For example, in the aviation industry, these algorithms are adopted for scheduling flights, deciding frequency of service to a particular location and crew allocation with the objective of profit maximization. Supply chain management is another field which relies heavily on optimization methods.

This course revisits the fundamental idea of “learning from interaction” as the theory behind one’s learning and intelligence. It elaborates the role of environment for natural learning to occur and thus the associated knowledge acquisition process. It is focused on goal-directed learning from interaction than the other approaches to machine learning. Concepts such as dynamic programming, temporal-difference, A3C, Actor-Critic, Deep-Q-Learning, Q_Sarsa, Value Function Approximation etc will be taught. The learners will be able to implement the techniques using PyTorch and will also be introduced to Mujoco and DeepMindLab.

This course introduces basic concepts in analysing time series data and is designed to illustrate the importance and applications of time series modelling. Topics covered include time-domain analysis, stationary processes, ACVF, ACF and PACF, Wolf decomposition, white noise process, cross-covariance, linear stationary processes, types of linear random processes.

This course also emphasizes solving case studies on different linear random processes and multivariate time series with exogenous input.