Karteikarten
AI Midterm
Karten in diesem Stapel (22)
- 1PhilosophyUnderstanding the nature of thought and the nature of intelligence
- 2PsychologyUnderstanding the mechanics of the human brain and its mental processes
- 3EngineeringUnderstanding how humans perform tasks and building algorithms to mimic this performance
- 4The early daysThe philosophy started even before computers, questioning what produced thought, and could an object have life. started with mechanical models. world war 2 motivated making computers that could do calculations people normally would. Alan Turing, John von-Nrumann, and claude shannon. big names
- 5The Turing Testderived from "Imitation Game". someone in a separate room would have to guess between a computer and person which is the human. This was done by asking questions, and trying to figurer it out based on response.
- 6The Chinese RoomWas made to prove the Turing Test meaningless. A computer took in Chinese characters, and produced other Chinese character as a response.
- 7Strong and weak AIThe Chinese room produced two ideas:
Strong AI - Hypothesis that machines would count as having actual minds.
Weak AI - Hypothesis that machines could possibly behave intelligently. - 8The Symbolic Eralate 1950s through early 80s.
Lead to the rule that the more knowledge you have, the less you search for the answer, but the better/faster you search, the less knowledge you need. used pathfinding, decision trees, state machines steering algorithms, and if... then. - 9The Modern Era1980s - early 90s due to two main issues:
symbolic worked with simple problems, but not complex ones, still couldn't mimic the human mind. move towards natural computing: neural network, genetic algorithms. - 10AI modelStrategy, Decision making, and movement
strategy is whole team based, where as decision making and movement are character-by-character. Some games may not need both group and character. It depends. - 11StrategyThe overall mindset a group has. Decision making will be influenced by group strategy, but each individual will have their own decision making.
- 12Decision MakingA single character working out what to do. Based on a range of different possible behaviours like attacking, idling... the decision making system picks the best choice. Usually based on proximity to things such as the player character.
- 13MomentTurns the decision making into some kind of motion. Can include path finding or other more complex algorithms to help navigate through environments.
- 14Agent based AIThe bottom-up design in working out how each character will behave and then implement the AI needed to produce that behaviour. an agent is anything that
- 15PEASPerformance, Environment, actuators, sensors.
- 16Simple reflex agentSelect action on the basis of current percept, ignoring all past percepts
- 17Model-based reflex agentmaintains a state that keeps track of parts of the environment that can't be observed at the moment
- 18Goal-based agentThis agent has information on it's goal, and picks the best possible action in the current state.
- 19Utility-based agentThe agent uses a utility function to evaluate the desirability of state that could result from each possible environment
- 20seekgets the direction of the target, and requests a velocity to take to said target.
- 21fleereverse of seek
steering.velocity = character.position - target.position - 22arrivelike seek, but with and added area around the target to stop, and start slowing down
Informationen
- Beschreibung
- Middle Game AI Karteikarten am AI Midterm, erstellt von Mackenzie Wilson am 20/10/2015.
- Erstellt von
- Mackenzie Wilson
- Aufrufe
- 3
- Bewertung