Map of the Field

Guide to AI and Data Courses
Purpose of the guideAI ⊃ ML ⊃ ANN ⊃ DLSix partsKinds of relationComparison criteriaInteractive map
Sercan KÜLCÜ · Giresun University · Computer Engineering

Introduction

What Is This Guide About?

Each of 38 courses on AI and data is introduced in one presentation. The aim is not to teach these courses but to map them: what question does each one ask, which tools does it share with others, and where does it part ways?

What you will find

  • Each course's core question and concepts
  • A short history and a typical workflow
  • Prerequisites and neighbouring courses
  • Overlap and difference tables
  • Applications and tools

What you will not find

  • Lecture notes or a textbook
  • Code examples or assignments
  • Derivations of formulas
  • The full syllabus of any course

Introduction

Nested Terms: AI ⊃ ML ⊃ ANN ⊃ DL

ArtificialIntelligencesearch, planning,logic,expert systems Machine Learningdecision trees,SVM, k-means Neural Networksperceptron,multilayer net Deep LearningCNN, transformer

Each ring is a subfield of the one around it: every deep learning model is a neural network, but not every AI system learns. A chess program that searches is AI, but it is not machine learning.

Structure

Six Parts, 38 Courses

Structure

Four Kinds of Relation Between Courses

subfield

Deep Learning → Neural Networks

Inherits all of its concepts and adds new ones of its own.

builds on

Computer Vision → Image Processing

The other course comes first; this one is built on top of it.

shares methods

Data Mining ↔ Machine Learning

They use the same tools but ask different questions.

application

Medical Image Processing → Computer Vision

Applies the other course's methods to one domain, under that domain's rules.

Structure

Criteria for Comparing Courses

Core question

The course's reason to exist, in one sentence.

Approach

The way of thinking it brings to a problem: search, statistics, logic, transforms…

Input

What kind of data or specification does the course start from?

Output

What comes out at the end: a decision, a model, an image, a report…

Success measure

How do we tell that a solution is good?

Math load

How much mathematics does it rely on, and which kind?

Data needs

How much data does it need; can it work without data?

Map

Interactive Map

Core
Branches of AI
Signals, Images and Audio
Data
Mathematical Foundations
Systems and Applications

Click a course: its neighbours light up and its relations are listed here.

Map

Links Between Parts

IIIIIIIVVVI
I. Core3132385
II. Branches of AI1379352
III. Signals, Images and Audio2910233
IV. Data332554
V. Mathematical Foundations85352·
VI. Systems and Applications5234·2
Each cell counts the relations between two parts; the diagonal shows links within a part. The core and the mathematical foundations connect to every part, while the systems courses rest mostly on the core.

Map

Hub Courses: The Most Connected

CourseLinksSome neighbours
Machine Learning15Artificial Intelligence, Neural Networks, Linear Algebra, Probability & Statistics …
Deep Learning9Neural Networks, GPU & Parallel Computing, NLP, Reinforcement Learning …
Artificial Intelligence8Machine Learning, Probability & Statistics, Optimization, NLP …
Computer Vision8Image Processing, Deep Learning, Computer Graphics, Pattern Recognition …
Data Mining8NLP, Machine Learning, Big Data, Probability & Statistics …
Probability & Statistics8Artificial Intelligence, Machine Learning, Pattern Recognition, Graphical Models …
NLP7Artificial Intelligence, Deep Learning, IR & Recommenders, Data Mining …
Linear Algebra7Machine Learning, Neural Networks, Computer Graphics, Robotics …
Machine learning and the mathematical foundations (linear algebra, probability, optimization) are the hubs of the map: most other courses rest on one of them.

How to Use

How to Read This Guide

Front to back

The core courses (Chapters 2–5) are the common language of the whole map; read them first.

From an interest

Images, language, data or robotics: start from the part that interests you and follow the prerequisite tables backwards.

By comparison

Open the "Where It Differs" slides of two courses you tend to confuse side by side.

Choosing electives

The roadmaps in the last chapter suggest groups of courses that make sense together.

The End

Chapter 1: Map of the Field
Next chapter: Chapter 2: Artificial Intelligence
1 / 1 Sercan KÜLCÜ. All rights reserved.