IIITH_Summerschool

Content presented in summer school on Machine Learning at IIIT Hyderabad from 09 - 15 July, 2017.

Day 0:

• Practice session on Python Basics, Intro. to Machine Learning and Pytorch basics

Day 1:

• Introduction to ML, Deep Learning, and Optimization by Prof. C.V. Jawahar (IIIT-H)
• Optimization and practical issues in Deep Learning by Prof. Vineeth Balasubramanian (IIT-H)
  • Backpropagation, Gradient Descent and its challenges
  • Algorithmic approaches for handling error surfaces (batch GD, stochastic GD, Mini batch SGD)
  • Methods to improve SGD - Momentum, Nestrov Accelerated Momentum
  • Learning rate (Adagrad, RMSprop, Adam, ...)
  • Advanced Optimization methods (Newton, Quasi-Newton, Conjugate Gradient, Hessian free optimization, Natural Gradient)
  • Regularization (Dropout)
  • Batch Normalization
  • Activation Functions

Day 2:

• Recurrent Neural networks by Prof Girish (IIIT-H) and Mr. Ankush Gupta (Ph.D. Student Uni. of Oxford)

Day 3:

• Introduction to Game theory by Prof. Sujit Prakash (IIIT-H)
• Generative adverserial networks - I, II by Prof. Vinay Namboodiri (IIT Kanpur)

Day 4:

• Variational Inference and Auto Encoders by Prof. Kaushik Mitra (IIT Madras)
• Variational auto encoders by prof. Vineeth Balasubramanian (IIT Hyd.)
• DARVIZ demonstration by Anush Sanakaran (IBM Bangalore)

Day 5:

• Reinforcement Learning by Prof. Ravindran Balaraman (IIT Madras)
• Deep Reinforcement Learning by Dr. Tejas Kulkarni (Google Deep Mind, London)

Day 6:

• Deep Learning for Language Processing by prof. Mithesh Khapra (IIT Madras)
• Model Compression by Prof. Girish Varma (IIIT Hyderabad)

---