Course Description
Deep Learning (or artificial neural networks) allows us to learn from data, rather than using rule-based software. Neural networks are a technique for building a computer program that learns from data. It is based very loosely on how we think the human brain works. First, a collection of software “neurons” are created and connected together, allowing them to send messages to each other. Next, the network is asked to solve a problem, which it attempts to do over and over, each time strengthening the connections that lead to success and diminishing those that lead to failure. TensorFlow has fast become the de facto framework for deep learning, after being open sourced by Google in November 2015. AI, which currently consists almost entirely of deep learning technologies, is predicted to become exponentially more prevalent over the coming years. Hence the importance of TensorFlow as the framework at the heart of this revolution. We at Ivy Data Science feel, therefore, that knowing TensorFlow well will prove to be extremely valuable in the years ahead.
What am I going to get from this course?
Comfortability with the TensorFlow command line and using TF on basic datasets to extract insights.
Prerequisites and Target Audience
What will students need to know or do before starting this course?
Students should be able to download and install software packages and be comfortable working in a Unix/Linux Shell environment.
Who should take this course? Who should not?
You should be comfortable with Python and Statistics.
Curriculum
Module 1: Day 1 - Deep Learning and TensorFlow Overview
Lecture 1
9am – 10am Introductions
Getting to know academic/professional backgrounds of candidates. Introducing the lecturer, course outline, and housekeeping.
Lecture 2
10am – 11am Introduction to Deep Learning
High level overview of the current deep learning landscape.
What is Deep Learning?
Why is it currently a popular machine learning framework?
Applications of Deep Learning.
Lecture 3
11:15am – Noon Code walk through
Download and install TensorFlow. Test the installation.
Lecture 5
1pm – 2pm Lunch
Lecture 6
2pm – 3pm Overview of TensorFlow
Data flow graph, Features, Functionality.
Lecture 7
3pm – 4pm TF Code walk through
TensorFlow at the command line, TensorBoard, NN Playground.
Lecture 8
4:15pm – 6pm Lab
Run TF from the command line, Run a demo, Train your first TensorFlow neural net model, Building the computation graph, Launching a graph session, Visualize your training using TensorBoard, Evaluating your model.
Module 2: Day 2 - Introduction to Convolutional Neural Networks
Lecture 9
9am – 10am Introduction to CNNs
Introduction to convolutional neural networks, CNNs in TensorFlow .
Lecture 10
10am – 11am Code walk through on MNIST data set
Identifying and downloading the data set, Training the model, Evaluating the model.
Lecture 11
11:15am – 1pm Lab
Load MNIST data set, Regression, Testing the model.
Lecture 12
1pm – 2pm Lunch
Lecture 13
2pm – 3pm More on CNNs
Lecture 14
3pm – 4pm Code walk through on MNIST data set
Lecture 15
4:15pm – 6pm Lab
Prepare the data, Build the graph, Build a multilayer Convolutional Network, Train and evaluate the model.
Module 3: Day 3 - More on Convolutional Neural Networks
Lecture 16
9am – 10am TensorFlow Serving
Architecture overview, Servers, Sources, Managers, Download and install TensorFlow Serving.
Lecture 17
10am – 11am Code walk through on TensorFlow Serving
Train and Export TensorFlow Model
Lecture 18
11:15am – 1pm Lab
Using MNIST data set, Train and Export TensorFlow Model, Serve the exported model, Test and run the Server.
Lecture 19
1pm – 2pm Lunch
Lecture 20
2pm – 3pm Further CNN
Further applications using CNNs.
Lecture 21
3pm – 4pm Code walk through on CIFAR-10 data set
Lecture 22
4:15pm – 6pm Lab
Identify and install the CIFAR-10 data set, Basic analysis on CIFAR-10 data set.
Module 4: Day 4 - Introduction to Recurrent Neural Networks
Lecture 23
9am – 10am Introduction to RNNs
Introduction to recurrent neural networks, History, Concepts, LSTM, RNNs in TensorFlow
Lecture 24
10am – 11am Code walk through on language data set
Identify and load Penn Tree Bank dataset, Preliminary analysis
Lecture 25
11:15am – 1pm Lab
Lecture 26
1pm – 2pm Lunch
Lecture 27
2pm – 3pm More on RNNs
Lecture 28
3pm – 4pm Code walk through on language data set
Penn Tree Bank dataset, Analysis
Lecture 29
4:15pm – 6pm Lab
Penn Tree Bank dataset, Further analysis
Module 5: Day 5 - More on Recurrent Neural Networks
Lecture 30
9am – 10am RNNs for machine translation
Lecture 31
10am – 11am Code walk through on second language data set
Machine translation data set
Lecture 32
11:15am – 1pm Lab
Download a machine translation data set, Begin translation
Lecture 33
1pm – 2pm Lunch
Lecture 34
2pm – 3pm More on RNNs
Lecture 35
3pm – 4pm Code walk through on second language data set
Lecture 36
4:15pm – 5:30pm Lab
Completion of machine translation exercise
Lecture 37
5:30pm- 6pm Wrap up