Introduction Flashcards
An AI program is called an ___ which has a certain goal
Can be used to determine the BEST SOLUTION given an INCOMPLETE INFORMATION
Intelligent agent
The agent interacts with the
Environment
Input from the environment which is sensed by the agent’s SENSORS are called
Percepts
It is the current state of the environment
Percept
Agent affects the environment using actions done by the
Actuators
Mapping of percepts to the actions
Control policy
Process of sensing and interacting with the environment
Perception action cycle
AI deals with ___
Formulation and or design of the CONTROL POLICY
Fields with application of AI
Medicine Robotics Web Finance Games
Property of an environment wherein what the agent can sense in any point in time is completely sufficient to make the best decision
Fully observable
Some elements of the environment is not known to the agent
Agent needs extra memory to store previous percepts
Partially observable
The action of the agent uniquely determines the outcome
Deterministic
Outcome may be affected by element of randomness
Stochastic
Environment does not change unless acted upon by the agent
Static
Environment may change while agent decides what to do
Dynamic
Number of possible states and actions are finite
Discrete
The number if possible states and actions are infinite
Continuous
Environment aims to defeat the agent
Adversarial
Environment is not there to defeat the agent
Benign
Most applicationfs for AI are full of __ which can make finding the optimal solution to a problem difficult to know
Uncertainty
Reasons for uncertainty
Limit if sensors Stochastic environment Adversaries Laziness Ignorance
