DA 225-O Deep Learning 3:1 (Summer 2022)

Course Instructor: : Deepak Subramani, CDS

Course description: This four-credit course will be offered every year in the Summer term as an elective course of the Data Science and Business Analytics (DSBA) M.Tech (Online) program. This course is aimed to be a graduate-level (200-series) course. The course introduces and trains students in different deep learning techniques to succeed in industry and research. At the end of the course, students would be able to design and use deep learning algorithms.

Syllabus

Deep Neural Networks, Backpropagation, Regularization and Optimizers; Deep Convolutional Neural Networks; Deep Sequence Modeling with Recurrent Architectures; Generative Modeling with Autoencoders and Generative Adversarial Networks (GANs); Deep Reinforcement Learning; Software for Deep Learning; Selected Applications and Case Studies from Computer Vision, Climate Analytics, Financial Analytics, Autonomous Vehicles.

Topics

Module 1: Deep Neural Networks: Building and Training Multilayer Perceptrons, Backpropagation, Gradient Problems, Regularization, Transfer Learning with pre-trained models

Module 2: Deep Convolutional Neural Networks: Convolutional Layers, Pooling Layers, CNN Architectures, Object Detection, Semantic Segmentation. Applications in Computer Vision and Remote Sensing.

Module 3: Deep Sequence Modeling with Recurrent Architectures: Recurrent Neurons and Layers, Training Recurrent Neural Networks, Forecasting a Time Series, Handling Long Sequences. Applications in NLP and Financial Analytics.

Module 4: Generative Modeling: Stacked Autoencoders, Convolutional Autoencoders, Recurrent Autoencoders, Variational Autoencoders, Generative Adversarial Networks

Module 5: Deep Reinforcement Learning: Policy Gradient Networks, Deep Q Networks. Application in Financial Analytics and Autonomous Underwater Vehicles.

Textbooks / References

  1. Aurelien Geron (2019). Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow. O’Reilly Media Inc.
  2. Goodfellow, I., Bengio, Y., & Courville, A. (2016). Deep learning. MIT press.
  3. Chollet, F. (2017). Deep learning with Python. Simon and Schuster.

Prerequisites: DA 224-O Practical Machine Learning or Equivalent

Grading:

  • Homework/Assignments 40%
  • Module quizzes 30%
  • Final exam 30%.