Visão Geral

Neste Curso Deep Learning com Keras & TensorFlow,  você aprenderá conceitos e modelos de aprendizado profundo usando os frameworks Keras e TensorFlow, para realizar algoritmos de aprendizado profundo.

Publico Alvo

• O Curso Deep Learning with Keras & TensorFlow  é destinado a engenheiros de software, cientistas de dados, analistas de dados e estatísticos com interesse em aprendizado profundo.

Pre-Requisitos

• Familiaridade com os fundamentos da programação, um bom entendimento dos fundamentos da estatística e da matemática e um bom entendimento dos conceitos de aprendizado de máquina.

Materiais

Conteúdo Programatico

Deep Learning with Tensor Flow

1. Welcome!
2. Learning Objectives 

Introduction to Tensorflow 

1. Introduction to TensorFlow
2. TF2x and Eager Execution
3. Tensorflow Hello World
4. Linear Regression With Tensorflow
5. Logistic Regression With Tensorflow
6. Intro to Deep Learning
7. Deep Neural Networks

Convolutional Networks

1. Intro to Convolutional Networks
2. CNN for Classifications
3. CNN Architecture
4. Understanding Convolutions
5. CNN with MNIST Dataset 

Recurrent Neural Network

1. The Sequential Problem
2. The RNN Model
3. The LSTM Model
4. LTSM Basics
5. Applying RNNs to Language Modeling
6. LSTM Language Modelling

Restricted Boltzmann Machines (RBM)

1. Intro to RBMs
2. Training RBMs
3. RBM with MNIST 

Autoencoders

1. Intro to Autoencoders
2. Autoencoder Structure
3. Autoencoders

Deep Learning with Keras and Tensor Flow

1. Course introduction
2. Introduction

AI and Deep learning introduction

1. What is AI and Deep learning
2. Brief History of AI
3. Recap: SL, UL and RL
4. Deep learning : successes last decade
5. Demo & discussion: Self driving car object detection
6. Applications of Deep learning
7. Challenges of Deep learning
8. Demo & discussion: Sentiment analysis using LSTM
9. Fullcycle of a deep learning project
10. Key Takeaways
11. Knowledge Check 

Artificial Neural Network

1. Biological Neuron Vs Perceptron
2. Shallow neural network
3. Training a Perceptron
4. Demo code: Perceptron ( linear classification) (Assisted)
5. Backpropagation
6. Role of Activation functions & backpropagation
7. Demo code: Backpropagation (Assisted)
8. Demo code: Activation Function (Unassisted)
9. Optimization
10. Regularization
11. Dropout layer
12. Key Takeaways
13. Knowledge Check
14. Lesson-end Project (MNIST Image Classification) 

Deep Neural Network & Tools

1. Deep Neural Network : why and applications
2. Designing a Deep neural network
3. How to choose your loss function?
4. Tools for Deep learning models
5. Keras and its Elements
6. Demo Code: Build a deep learning model using Keras (Assisted)
7. Tensorflow and Its ecosystem
8. Demo Code: Build a deep learning model using Tensorflow (Assisted)
9. TFlearn
10. Pytorch and its elements

Deep Neural Net optimization, tuning, interpretability

1. Optimization algorithms
2. SGD, Momentum, NAG, Adagrad, Adadelta , RMSprop, Adam
3. Batch normalization
4. Demo Code: Batch Normalization (Assisted)
5. Exploding and vanishing gradients
6. Hyperparameter tuning
7. Interpretability

Convolutional Neural Network

1. Success and history
2. CNN Network design and architecture
3. Demo code: CNN Image Classification (Assisted)
4. Deep convolutional models 

Recurrent Neural Networks

1. Sequence data
2. Sense of time
3. RNN introduction
4. LSTM ( retail sales dataset kaggle)
5. Demo code: Stock Price Prediction with LSTM (Assisted)
6. Demo code: Multiclass Classification using LSTM (Unassisted)
7. Demo code: Sentiment Analysis using LSTM (Assisted)
8. GRUs
9. LSTM Vs GRUs

Autoencoders

1. Introduction to Autoencoders
2. Applications of Autoencoders
3. Autoencoder for anomaly detection
4. Demo code: Autoencoder model for MNIST data (Assisted)