01.3/4 Rationality & AI Paradigms

Rational agent

takes actions it believes will achieve its goals

Limited Rationality

acting the best way you can subject to the computational constraints that you have

Limited Rationality

we need to program the agent’s head that’s going to last for the agent’s whole range of things it has to do and life it has to live

Sequence of Percepts

learning, part of the agent’s job is to use _ _ _ to estimate the missing details in the world dynamics

Percept, decision, action, environment responds

agent loop

Rational agent

maximizes performance measure

Fully observable vs. Partially observable

environment types, categories (1)

Deterministic vs. Stochastic

environment types, categories (2)

Episodic vs. Sequential

environment types, categories (3)

Static vs. Dynamic

environment types, categories (4)

Discrete vs. Continuous

environment types, categories (5)

Paradigms

they shape how agents think

Paradigms

they decide how “smart” looks in context

Accessible vs. Inaccessible

can you see the state of the world directly?

Deterministic vs. Non-deterministic

does an action map one state into a single other state

State vs. Dynamic

can the world change while you are thinking

Discrete s. Continuous

are the percepts and actions discrete (like integers) or continuous (like reals)