Artificial Intelligence

Artificial Intelligence is a kind of intelligence which can solve a problem. A system is “intelligent” if a person cannot distinguish its behaviour from a human’s.

Link to original

Turing Test

The Turing test instructs:

  • Let a human ask questions to a system.
  • If the human cannot tell whether the answers were given by a human or machine, it passes the test.

The Total Turing test expands on this and says a robot is intelligent if a person cannot distinguish its behaviour from a human’s, including physical behaviour.

Link to original

Model-based AI approach

Model-based AI approach, e.g. graph and cost model:

  • interpret-able (make sense)
  • many search techniques exploit models to arrive at a solution quickly
Link to original

Data-driven AI approach

Data-driven AI approach, e.g. deep learning:

  • uses a generic model
  • no domain knowledge needed
  • may require lots of data / computing power
  • can learn complex models
Link to original

Applications of search:

  • Mapping
  • Computer Games
  • Robotics
  • Network routing
  • Scheduling (calendar, CPU)
  • Playlist generation

AI Agents

AI Agent

An agent is a unit that makes an intelligent decision given the opportunity, it uses the knowledge it has at that point in time to make a decision. Decision is limited by range of actions possible at that point in time.

Link to original

Autonomous Agent

An autonomous agent makes decisions without human input.

Link to original

Rational Agent

A rational agent makes decisions based on knowledge it has (agent acting in the real world under resource constraints).

An ideal rational agent will act to “maximise its expected performance measure” on the basis of provided information by ‘percepts’ from the environment. However rational doesn’t mean perfect or omniscient.

Link to original