Module 1: What is AI?
Topic 1: Defining AI.
No brain, no train, no AI
- 4 common characteristics:
1. Seeing and understanding the complex world around us,
2. Handling information: collecting and understanding information,
3. Making choices: learning from information and deciding what to do; and
4. Working to achieve tasks or goals.
Topic 2: Different levels of AI Definitions
a. The abstract level
Create machines that are able to think and solve problems like humans
b. The legal level
AI system’ means a machine-based system that is designed to operate with varying levels of
autonomy and that may exhibit adaptiveness after deployment, and that, for explicit or implicit
objectives, infers, from the input it receives, how to generate outputs such as predictions, content,
recommendations, or decisions that can influence physical or virtual environments.
=> Work on its own (autonomy) & learn and change (adaptiveness) -> dif from other technologies
that already have legal rules.
=> Ability: make predictions or decisions.
* Important for making rules about AI because it helps us to understand which specific systems the
rules should cover.
c. The technical level
Artificial intelligence (AI) is a branch of computer science that studies the computational
requirements for tasks such as perception, reasoning, and learning, and develops systems to perform
those tasks.
It was trained from past information, knowledge or experience to solve problems. From a computer
science perspective, this means the developers had to define the computational tasks the system
should perform. For example, the task is prediction. The system "learns" by analysing lots of data and
adjusting its internal mathematical model until it becomes good at making these predictions.
Topic 3: AI, Machine Learning and Deep Learning
a. Machine learning
The most widely used technological approach
Machine learning is a part of artificial intelligence that looks at how computers can get better at
understanding, thinking, or acting by learning from experience or data. It uses ideas from computer
science, math, psychology, brain science, economics, and other areas to do this.
- Two classes of ML:
+ Discriminative models or predictive AI:
Solve yes-or-no problems
Sort things
Eg: a spam-filter or a tool that names pictures.
+ Generative models or generative AI:
Make text or pictures => Create
Eg: While models that tell things apart learn a pattern to give correct labels, for example to
spot a husky or a horse in a picture, models that create things "learn" those patterns to make
their own pictures of huskies and horses.
* Always adapt which is what makes them different from simple AI.
b. Deep learning
The most popular type of machine learning today is called deep learning.
Deep learning is a type of machine learning that uses large networks of artificial "neurons" to process
information.
The networks that are used in deep learning are called neural networks. They are like a way to
organise how machines learn. they do three main things:
1. They get information
2. They process this information
3. They give an answer or result
We use the word "deep" because the information goes through many steps when the machine is
thinking about it. These steps are called layers.