Mind map
Agents

Mind map: agents artificial intelligence
Map content
Agents
Simple reflex agents
- Actions selected depend on solely the current state (Ignores percept history)
- Has to be able to observe the whole environment
Model-based reflex agents
- Tracks the environment even when observing it (Percept History)
- Aware of:
- How things around it evolve (independent of it).
- How it influences the world by its actions
Goal-based agents = Model-based but with objectives
- Searching
- Planning
- Considers future
Utility-based agents
- Compare different worlds using utility function
- Map state(s) with numbers associated with happiness
- Applies tradeoff to conflicting goals
- Weighs up likelyhood of sucess & importance of task in uncertain goals
- Compare different worlds using utility function
Learning agents
- Makes improvements using feedback
- analogy: perception split as reward/punishment
- Suggests actions that lead to informative experience
- Makes improvements using feedback