infyni

Artificial Intelligence with a focus on Deep Learning - Today’s in-demand Skillset

Deep learning is a subset of machine learning in artificial intelligence (AI) that has networks capable of learning unsupervised from data that is unstructured or unlabeled.

Live Course

Duration: 24 Hours

Enrolled: 1

Offered by: infyni

Live Course

About Course

Start your Artificial Intelligence and Machine Learning journey by joining “Deep Learning and its applications : Beginners to Advance” course. The program builds a solid foundation from basics to advance by covering the most popular and widely used deep learning technologies and its applications. The course is divided into six modules where each module will be covered in 3 Hrs theory sessions along with 1 Hr of practical hands-on session. Hands-on lab sessions will be covered on AWS Cloud platform & using Jupyter Notebook
The course content includes six fundamental topics viz.
(1) Basics and Fully Connected Networks.
(2) Convolutional Neural Networks (CNN),
(3) Object Detection (RCNN, SSD, YOLO),
(4) Recurrent Neural Networks (RNN, LSTM, GRU),
(5) Auto-Encoder & VAE and finally
(6) Generative Adversarial Networks.

Topics
Instructor

1 Introduction and Fully Connected Network

Introduction: Basic classification and regression
Machine learning vs Deep Learning
Basics of neural networks, single and multi-layer perception algorithm
Intuition behind Fully connected networks: Single and multi-layered

2 Introduction and Fully Connected Network (Contd...)

Loss functions
backpropagation
computational graphs
optimizers
Network initialization
Activation functions
Batch normalization
One hour hands-on practical session

3 Convolution Neural Networks - CNN

Pros and cons of Fully connected networks
Weight sharing in CNN
CNN intuition
CNN specifications and parameterizations
Max pooling, upsampling

4 Convolution Neural Networks - CNN (Contd...)

Case studies- AlexNet, VGS
Inception
ResNet
DenseNet etc
One hour hands-on practical session.

5 Designing Object detection Networks - RNN

Problem definition and its solution via. Regression and Classification
Overfeat network for object detection
Region CNN (R-CNN)
Fast and Faster RCNN object detection networks
You Look only once (YOLO) based object detection.

6 Designing Object detection Networks - RNN (Contd...)

Single Shot Detection(SSD) based object detection
One hour hands-on practical session.

7 Recurrent Neural Networks - RNN

Pros and cons of CNN and Time series data.
Intuition of Recurrent Neural Networks (RNN).
Training and testing of recurrent units, Backpropagation in time
Problem of Vanishing and exploding gradients.

8 Recurrent Neural Networks - RNN (Contd...)

Long term and short term memory (LSTM) and Gated recurrent units (GRU).
One hour hands-on practical session.

9 Auto-Encoders and Variational Auto-Encoders

Basic intuition and motivation of Auto-encoders and their requirements.
Intuition behind Variational Auto-encoders and Image generation using VAE

10 Auto-Encoders and Variational Auto-Encoders (Contd...)

Theoretical background of VAE
Pros and cons of VAE
One hour hands-on practical session.

11 Generative Adversarial Networks

Intuition behind Generative Adversarial Networks (GAN)
Theoretical background of GAN
GAN practically

12 Generative Adversarial Networks (Contd...)

DC GAN, C-GAN
STACK-GAN
Pix2Pix
CycleGAN
One hour hands-on practical session