Tag: Advanced Deep Learning

Duration

14 hours (usually 2 days including breaks)

Requirements

• An understanding of basic linear algebra

Audience

• Software Engineers

Overview

Keras is an open source Python neural-network library for the creation of deep learning neural-networks. Keras offers an intuitive set of abstractions, simplifying the development of deep learning neural-networks and models.

This instructor-led, live training (online or onsite) is aimed at software engineers who wish to develop advanced deep learning neural-networks and model using Keras and Python.

By the end of this training, participants will be able to:

• Apply deep learning with supervised or unsupervised learning methods.
• Develop, train, and implement concurrent neural networks and recurrent neural networks.
• Use Keras and Python to build deep learning models to solve problems involving images, text, sound, and more.

Format of the Course

• Interactive lecture and discussion.
• Lots of exercises and practice.
• Hands-on implementation in a live-lab environment.

Course Customization Options

• To request a customized training for this course, please contact us to arrange.

Course Outline

Introduction

Keras and Deep Learning Frameworks

• TensorFlow and Theano back-ends
• Keras vs Tensorflow

Data and Machine Learning

• Tabular data, visual data, unstructured data, etc.
• Unsupervised learning, supervised learning, reinforcement learning, etc.

Preparing the Development Environment

• Installing and configuring Anaconda
• Installing Keras with a TensorFlow back-end

Neural Networks in Keras

• Using Keras functional API to build a network
• Pre-processing and fitting data
• Defining a Keras model

Mutiple Input and Output Networks

• Building two input-networks
• Representing high-cardinality data
• Merging layers
• Extending the two input-network
• Building neural networks with multiple outputs
• Solving multiple problems simultaneously

Training and Pre-Training

• Training models
• Saving and loading models
• Using ResNet50 on models

TensorBoard

• Exporting Keras logs
• Visualizing a computational graph and training progress

Google Cloud

• Exporting models
• Uploading Keras models
• Using a model in Google Cloud

Summary and Conclusion

Duration

28 hours (usually 4 days including breaks)

Requirements

Good understanding of Machine Learning. At least theoretical knowledge of Deep Learning.

Overview

Machine learning is a branch of Artificial Intelligence wherein computers have the ability to learn without being explicitly programmed. Deep learning is a subfield of machine learning which uses methods based on learning data representations and structures such as neural networks.

Course Outline

• Machine Learning Limitations
• Machine Learning, Non-linear mappings
• Neural Networks
• Non-Linear Optimization, Stochastic/MiniBatch Gradient Decent
• Back Propagation
• Deep Sparse Coding
• Sparse Autoencoders (SAE)
• Convolutional Neural Networks (CNNs)
• Successes: Descriptor Matching
• Stereo-based Obstacle
• Avoidance for Robotics
• Pooling and invariance
• Visualization/Deconvolutional Networks
• Recurrent Neural Networks (RNNs) and their optimizaiton
• Applications to NLP
• RNNs continued,
• Hessian-Free Optimization
• Language analysis: word/sentence vectors, parsing, sentiment analysis, etc.
• Probabilistic Graphical Models
• Hopfield Nets, Boltzmann machines
• Deep Belief Nets, Stacked RBMs
• Applications to NLP, Pose and Activity Recognition in Videos
• Recent Advances
• Large-Scale Learning
• Neural Turing Machines