Scribd
Upload
2 views

Lecture 1 Logistic.pptx

The document outlines the course structure and logistics for a Deep Learning class (DS5007) taught by Sahely Bhadra, including lecture and lab timings, prerequisites, evaluation criteria, and lab assignments. It emphasizes the importance of understanding neural networks, their applications, and the factors contributing to the success of deep learning. The course aims to equip students with the knowledge to build deep learning models and read relevant research papers, while cautioning that expertise requires more than just one course.

Uploaded by

abby.iitpkd
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
2 views

Lecture 1 Logistic.pptx

The document outlines the course structure and logistics for a Deep Learning class (DS5007) taught by Sahely Bhadra, including lecture and lab timings, prerequisites, evaluation criteria, and lab assignments. It emphasizes the importance of understanding neural networks, their applications, and the factors contributing to the success of deep learning. The course aims to equip students with the knowledge to build deep learning models and read relevant research papers, while cautioning that expertise requires more than just one course.

Uploaded by

abby.iitpkd
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 19

Deep Learning

DS5007
Sahely Bhadra
[email protected]
Logistics and Other Trivia
• Course Structure – 3-0-2-4 • Office Hours
• Lecture Timings (proposed) • Instructor – by appointment
• Tuesday – 2:00-3:15 pm
• Thursday - 1:45-3:00 pm
• Lab Hours
• Friday - 10:00am-12:50pm
• TA –
• Vishnu
• Arjun
• Bhaskar

2
Reference Material
• No fixed textbook, primary reference material (all three are
available online freely!)

• Lots and lots of research papers

3
Prerequisites
• Officially – Machine Learning
• Calculus
• Linear algebra
• Probability and Statistics
• Exposure to machine learning concepts will be helpful.
• If you are doubtful, please talk to the instructor

4
Evaluation criteria
• Test 1: 10
• Test 2: 10
• End Sem: 40
• Q and A: 10
• Programming assignment: 15(6 assignment)
• Project: 15 (novel component essential (10 +5))

5
Labs Lab Due

L0 Jan10
• Programming Assignments
L1 Jan 24
• Start early, experiments will taking time to execute
L2 Feb 7
• For project can be done in a teams of 2
Mid Feb14
• L0 – warm up/practice lab to get familiarized with Proj
the framework L3 Feb21

• L1 to L6 you have to give a viva in lab. L4 Mar07

• Encourage GitHub repositories of your L5 Mar21


submission, along with blogs explaining your L6 Apr4
findings Final Apr11

6
Deep Learning aka Artificial Neural
Networks

7
Deep Learning Success
• Computer Vision
• Object Recognition

8
Deep Learning Success
• Computer Vision
• Object Recognition

9
Deep Learning Success
• Computer Vision
• Object Recognition
• Automated Image Captioning

10
Deep Learning Success
• Computer Vision
• Object Recognition
• Automated Image Captioning
• Autonomous Driving

11
Deep Learning Success
• Computer Vision
• Object Recognition
• Automated Image Captioning
• Autonomous Driving
• Game Playing

12
Deep Learning Success
• Computer Vision
• Object Recognition
• Automated Image Captioning
• Autonomous Driving
• Game Playing
• Natural Language Processing

13
Deep Learning Success
• Computer Vision
• Object Recognition
• Automated Image Captioning
• Autonomous Driving
• Game Playing
• Natural Language Processing
• Speech Processing

14
Deep Learning Success
• Computer Vision
• Object Recognition
• Automated Image Captioning
• Autonomous Driving
• Game Playing
• Natural Language Processing
• Speech Processing
• Creativity

15
Deep Learning Success
• Computer Vision
• Object Recognition
• Automated Image Captioning
• Autonomous Driving
• Game Playing
• Natural Language Processing
• Speech Processing
• Creativity
• And a whole lot of others…
16
Deep
Learning vs
Traditional
ML

17
Factors Contributing to the Success of
DL
• Increasing Size of Datasets
• Increasing Model Sizes
• Increasing Compute Power

18
Objectives of This Course
• Understanding Neural Networks aka Deep Learning
• Principles and Internals of the models
• Familiarity with the terminology
• Ability to build them from scratch for various applications
• Train the networks patiently!
• Equip with the knowledge required to read research papers on
the topic
• Caution: You will not become an expert in the field just because
you did one course!

19

You might also like