IB ITGS HL - 16. AI & Robotics

What is intelligence?

Intelligence is an abstract concept. It can be defined as the ability to:
- Respond to the environment
- Learn new knowledge
- Use logic to come to a conclusion
- Learn from experience
- Make evaluations or judgments
Artificial Intelligence refers to the systems that simulate intelligence through a series of rigid facts or rules.
Computational Intelligence focuses on creating systems that think in the same way that humans do.

The experiment involves a human judge who asks questions to two contestants.
The judge tries to determine which of the contestants is human and which one is a computer.
If the judge is unable to tell the difference, the computer is deemed “intelligent”

Modern form of the Turing Test.
Designed to prevent spam bots advertising in comments or creating false accounts.
It displays a deformed text, which is only legible to a human.

Expert systems are software programs that use programmed logic and rules to make the same decisions as a human expert.
They are usually restricted to answering questions in a specific knowledge domain rather than being general experts on everything.

Expert system shell: set of programs which allow the building of an expert system through the creation of knowledge and rules.
User interface: presents questions to the user and accepts inputs from them.
Knowledge base: contains data and facts which form the knowledge in the knowledge domain.
Knowledge engineer: knowledge from the expert is given to the system.
Inference engine: has the job of matching the user’s input with the data contained in the knowledge base. This is done using inference rules.

Fuzzy logic is used to model concepts that don’t have a concrete yes/no answer.
Allows to provide answers with differing degrees of certainty.

Considers most, if not all, possible solutions to a problem in order to find the most appropriate one.

Natural language processing refers to the ability of a computer to understand human languages. It may be used to take input from the user.
Machine translation: translating from one human language to another.
- Grammar can be an issue, because sentences are formatted differently in different languages.

Knowledge can be represented using nodes that are connected by links of varying weights.

Artificial neural networks are an attempt to make computers learn in a similar way to humans.
- It needs training before it can be used.
- Each node takes inputs from several input nodes and uses a transfer function to determine its output.
- Back propagation: the weights of some nodes in the ANN are adjusted until the transfer function produce output more closely matched to the expected output.
- They feature a series of inputs and outputs, connected by one or more layers of nodes.

A robot is a computer-controlled system that performs manual, physical tasks
- They can be autonomous by using CI or AI.
- Or they can be remotely controlled by a human operator.
They perform three types of physical jobs:
- Dangerous jobs
- Boring or repetitive jobs
- Exploring inaccessible environments