DEEP LEARNING
HOME Schedule Students List Assignments Contact-CB
 
 

This elective course introduces students to the fundamentals of deep learning and takes them on a journey through the whole zoo of architectures, design and applications. Emphasis is on learning concepts related to the underlying principles of machine learning in general. The course is structured in a goal-oriented fashion with a well-known application forming the skeleton: in this case, the Telugu OCR system developed in the School. Students will hack their way through the OCR system while learning the DL concepts.

Textbook(s):

In addition, several technical papers, videos and other material will be provided as needed.

There are some fun, and more or less informal, notes on machine learning. Haven't yet proof-read them thoroughly; check and report mistakes. Thanks!

Class Timings

Mondays: 11:00 AM - 1:00 PM
Tuesdays: 12:00 PM - 1:00 PM

Instructors
Srinivasa Rao, B.
Chakravarthy Bhagvati

 

Background

Image Processing, Pattern Recognition

The above will be covered in the first few classes but students must do self-study based on web tutorials.

 
 
 
 
 
 

List of Projects (Tentative)

The following is a possible list of projects for you to choose from. We will finalise the list by 5 March.

  1. Image Captioning Using Deep Transfer Learning
  2. Vehicle Over Speed Detection System
  3. Deep Learning-Based Conjunctival Melanoma Detection Using Ocular Surface Images
  4. Automatic detection of pathological changes in chest X-ray screening images using deep learning methods
  5. Colourising grayscale images
  6. Style Transfer in Fashion Industry using GANs
  7. Deep learning-based multiobject tracking
  8. Voice cloning
  9. Deep Learning Model for Analysis of Plant Diseases
  10. Speech Emotion Recognition Using Deep CNNs
  11. Image deblurring using Deep CNNs
  12. Audio from Lip-syncing videos
  13. Converting handwritten to printed characters and vice-versa
  14. VAE based sentence interpolation in Telugu

Colah's blog on LSTMs

 

Andrej Karpathy's blog on RNNs

 
 

Contact

Chakravarthy Bhagvati
Professor
School of Computer and Information Sciences
AI Lab, University of Hyderabad
Hyderabad - 500046.

Note that my preferred contact is via email unless it is an emergency (should be very, very rare!). I usually respond within a day.

My mobile number is with the class representative and a text message may be sent in certain cases where email is not appropriate.

Call me only as a last resort!


Home Page

 

Email: chakcs@@uohyd..ernet..in
(Remove duplicate symbols!)