/dscamp_public

Materials for Data Science camp

Primary LanguageJupyter Notebook

Data Science Camp materials

This repository contains code and materials for Data Science camp. Please see outline of the modules below. To learn about specific module and access its learning materials, navigate to the corresponding directory.

Outline

Day Week 1 Week 2
Monday Intro to Data Science/ Motivating examples for each session/Set-up in Google Colab 2nd NLP session and Transformer
Tuesday Python Programming/Q&A for Google Colab Set-up Reinforcement learning (RL)
Wednesday Intro to Models (Linear Models, Random Numbers and Distributions) GAN – Image Manipulation
Thursday Neural Networks Preparation to AI Startup Team Project and Team Neural Network Task
Friday Intro to AI Startup Team Project/Neural Network Team Task/Intro to NLP (1st session) Presentations of AI Startup and Team NN Task; Wrap-up

Modules

Each of the modules is designed to be a self-contained units, with themes that are connected across the models.

Introduction to Data Science (1 unit)

Motivating Example(s): Many AI in the world today;

Topics: -- Introduce Data Science and AI; Examples from the modern world; Give highlight/overview of project;

Objectives: -- Learning from Data w/Computers

Project(s): No Project

Curriculum Threads: What is a model; show how do computers fit it

Materials: Introductory material; motivating examples

Introduction to Python

Motivating Example(s): Lint Code

Topics: Starting and Using Jupyter Notebooks; Assignment, Control Loops, Data Structures

Objectives: -- Introduce basic syntax of Python programming language and its ecosystem;

Project(s): A competition on Lint Code

Curriculum Threads: Optimization -- Computer does it for you

Materials: Pet examples from Lint Code

Linear Models and Random Numbers (1 unit)

Motivating Example(s): Scientific papers that use regressions in its analysis

Topics: Conceptual introduction of models (linear and logistic regression), random numbers (distributions)

Objectives: Introduce concepts of linear statistical modelling and random numbers

Project(s): Social network model of credit decisioning using logistic regression

Curriculum Threads: It starts with drawing a line through the data points you have

Materials: Linear and Logistic Regression materials, Random Numbers materials

Artificial Neural Networks and Image Recognition (1 unit)

Motivating Example(s): Photo Classification on the Social Network

Topics: Biological and Artificial Neural Networks, Splines and Nonlinear Regression, Artificial Neural Network Structure(s); Convolutional Networks

Objectives: At the conclusion of the unit, students will be able to

  • Describe the relationships between artificial neurons and biological neurons
  • Build simple neural networks using TensorFlow
  • Build a neural network to classify images using TensorFlow
  • Use a Neural Network model to build an object detector.

Project(s): Object Detector

Curriculum Threads: Modeling (Biological model vs Mathematical), Optimization (Gradient Descent)

Natural Language Processing (2 units)

Motivating Example(s): Email completion, Language translator, Text Generation (tweets, Shakespeare, Reviews, chat-up lines)

Topics: Text Representation; Recurrent Neural Networks, Convolutional Neural Networks

Objectives: The main topics to be discussed in this part are:

  • Brief introduction to the world of natural language processing (NLP) and show students some everyday applications of NLP.
  • Help students develop simple NLP projects and how the projects are used in large scale.
  • Motivate student by engage students to interact with AI.
  • Two activities include writing essays using AI and build a chatbot.

Project(s): Sentiment Classification, Language Generation, Chatbot (s)

Curriculum Threads: Natural language processing, Deep learning models

Reinforcement Learning (1 unit)

Motivating Example(s): Mar I/O (RL agent playing Super Mario)

Topics: Reinforcement Learning and dynamic programming

Objectives: Gentle introduction to reinforcement learning concepts

Project(s): Procgen

Curriculum Threads: RL; Q-learning; Markov Chains

GAN (1 unit)

Motivating Example(s): Handwriting Recognition

Topics: GAN structure

Objectives: Have a basic understanding of GAN and distrbution generation

Project(s): Generate Anime Faces

Curriculum Threads: Models; Optimization

Project: AI Startup (1 unit)

Motivating Example(s): Applications from the course of the week

Topics: Applications of Machine Learning and AI

Objectives: Teams of students design an AI startup company, using the principles of AI and machine learning from the course of the camp.

Project(s): Student-generated startup "pitches" to convince a group of venture capitalists to invest in their startup.

Curriculum Threads: Applications of AI

Presentation and Wrap-Up (1 unit)

Topics: Applications of Machine Learning and AI, Careers in AI

Objectives:

  • Students give their pitches to a panel of "investors"/judges. The top group may win an edible prize, if available.
  • A panel of instructors describe their personal journeys to become data scientists, giving students a picture of the diverse options ahead of them.
  • Open Question and Answers about Careers and AI careers
  • Information on additional resources and how to continue learning

Project(s): Student-generated startup "pitches" to convince a group of venture capitalists to invest in their startup.