Online Deep Learning Course For Humanists

Course Outline

Background Knowledge

Programming

• Python Basic

Math

• basic matrix, calculus, and statistics.

Chapter 1 Introduction

1.1 Introduction

1. What is Artificial Intelligence?
2. What is Machine Learning
3. What is Deep Learning? Why we use it?
4. Relationship: AI-> machine learning -> deep learning

1.2 Environment Setup

Setup the environment we need in this course, including Anaconda, TensorFlow and Jupyter Lab.

Chapter 2 Perceptron

2.1 What is Perceptron?

1. Neuron and Perception

• Biological neuron model
• Artificial neural
• Mathematics form

2.2 Iris Classification

• Question and data set
• Linear classifier
• Implement a perceptron

2.3 Multilayer Perceptron

• XOR question

Chapter 3 Neural Networks

3.1 Two layer Neural Network

The architecture:

• Nodes
• Input/Output
• Layer
  • Input Layer
  • Output Layer
  • Hidden Layer: Why we call it hidden layer?
• Connection
  • Fully connected
• Weights

Activation function

• Why need this? None-linearities
  • Some common activation functions

3.2 Forward Propagation

3.3 Backward Propagation

• Math: Matrix Multiplication

Make an animation videos.

Chapter 4. Learning from data: Training Neural Networks

4.1 Example

• Error loss The empirical loss measures the total loss over the dataset

4.2 Loss Optimization

• Loss function is a function of the Weight

4.3 Gradient Descent

• Greedy algorithm
• Like Hiking Down a Mountain Make an animation video
• Local minimum

Make an animation videos.

4.4 Optimization

Physical Experiment

• Every student acted as a Neuron
• Mock Forward Propagation
• Mock Backward Propagation

Chapter 5 Make your own neural network to classify handwritten digits

In this chapter, the student will learn how to teach the computer to classify handwritten digits by using MNIST dataset in Python.

DataSet: The dataset I choose for this part is MNIST(Modified National Institute of Standards and Technology) dataset, which has a training set of 60,000 examples, and a test set of 10,000 examples. It is a subset of a larger set available from NIST(National Institute of Standards and Technology) which gives data set of over 800,000 images of handwritten digits from 3,600 writers. The digits have been size-normalized and centered in a fixed-size image.

Experiment

Run Chapter 5's code on Raspberry Pi.

Project

NLP Project using the red hen lab's News Dataset

Classify Hand Gesture Pose in Art

In this part, the student will learn how to classify Hand Gesture Pose in Art using Tensorflow.

Dataset:we will use Rijksmuseum dataset, the Fototeca Zeri dataor the Bibliotheca Hertziana dataset.

Audio Recognition of Simple words

In this part, the student will learn how to build a basic speech recognition network that recognizes ten different words: "yes", "no", "up", "down", "left", "right", "on", "off", "stop", or "go". Real speech and audio recognition systems is much more complex than classify handwritten digits.

Dataset: we will use the Speech Commands dataset, which consists of over 105,000 WAVE audio files of people saying thirty different words. This data was collected by Google and released under a CC-BY license.

Exploring Sentiment in Literature

In this part, the student will learn how to build use Convolutional Neural Network and Recurrent Neural Network to train a models, which will let us exploring Sentiment in Literature.

Dataset: we will choose three books (one positive book, one negative book, and one normal book) form Project Gutenberg. Project Gutenberg offers over 58,000 free eBooks.