Description:
View deep learning as a computational graph constructed from a group of function approximators. Develop deep neural networks to solve classification and regression tasks in computer vision and natural language processing.
Prerequisites:
The courses assignments and notes will use python programming language and expects a basic knowledge of python. We assume the student has completed the mathematics section of the Machine Learning Fundamentals course or has an equivalent fluency in mathematics and fundamentals.
Note: If you haven’t done the Machine learning fundamentals you will have to pass a test that verifies your understanding of fundamentals.
Schedule
Duration: 4 Weeks
|
Week |
Date |
Topic |
|
1 |
12-August |
Artificial Neural Networks |
|
1 |
13-August |
Backpropagation and loss |
|
2 |
19-August |
Deep Neural Networks |
|
2 |
20-August |
Break |
|
3 |
26-August |
Convolutional Networks |
|
3 |
27-August |
Recurrent Neural Networks |
|
4 |
02-September |
Transfer Learning |
|
5 |
03-September |
Syllabus
- Artificial Neural Networks:- Develop artificial neural networks as a graph computation and look at them from the lense of a programming pipeline.
- Backpropagation and Loss:- Dive deeper into understanding loss metrics and computing loss gradients across a computational graph.
- Deep Neural Networks:- Accomplish more complex tasks using deep neural networks and investigate optimization of deep networks.
- Convolutional Networks:- Transition from engineering features to learning features using deep convolutional neural networks for image recognition.
- Recurrent Networks:- Construct sequence based models using recurrent neural networks to arrive at solutions for simple language based tasks.
- Transfer Learning:- Understand transferring learned experiences across models to enhance learning using deep neural networks.
Cost: INR 50,000