Agents and Environments Flashcards
What does an agent do?
perceives its environment through sensors and acts upon it through actuators
What is a rational agent?
Acting rationally
- Chooses actions to maximize the expected utility using logical reasoning
- Rational to explore and learn
What is a Turing Test?
Acting humanly
A computer and human are asked questions to determine which is human. If the computer is indistinguishable from the human, it passes the Turing Test.
What is cognitive modeling?
Thinking humanly
Simulating human problem-solving and mental processing in a computerized model
What is formal logic?
Thinking rationally
The validation behind the thought in artificial intelligence
What is the gorilla problem
It is not obvious that we can control machines that are more intelligent than us.
What are the environment types for different agent designs?
- Fully/partially observable = agent requires memory
- Discrete/continuous = agent may not be able to enumerate all states
- Stochastic/deterministic = agent may have to prepare for contingencies
- Single-agent/multi-agent = agent may need to behave randomly
What is a reflex agent?
Chooses actions based on current percept
May have memory or a model of the world’s current state
Do NOT consider the future consequences of their actions
They can be rational
What is a planning agent?
Ask “what if”
Decisions are based on consequences of actions
Must have a model of how the world evolves in response to actions
Considers how the world WOULD BE
Optimal vs. complete planning
Optimal planning gives the best solution to a given problem
Complete planning considers a problem once all possible solutions are explored
What are the parts of a search problem and solution?
Search problems consist of: a state space, successor function (with actions, costs), start state, and goal test
Solution is a sequence of actions (a plan) which transforms the start state to a goal state
What is Depth-First Search?
Expand the deepest node first (fringe is LIFO stack)
Not optimal
What is Breadth-First Search?
Expand the shallowest node first (fringe is FIFO queue)
Optimal only if costs are all 1
When will BFS outperform DFS?
On sparse graphs with low branching factors, where DFS can get stuck following long, irrelevant chains.
When will DFS outperform BFS?
On thickly connected graphs with high branching factors, because BFS experiences an exponential blowup in nodes expanded as the branching factor increases.
What is a Uniform Cost Search?
Expand the cheapest node first (fringe is priority queue)
- Complete and optimal, though explores options in every direction
- no info on goal location
What does an agent function do?
Maps from the set of all possible percept sequences P* to the set of actions A formulated as an abstract mathematical function
f: P* -> A
What is PEAS?
Performance measure, Environment, Actuators, Sensors
- Crucial in AI design for tasks; serves as a framework for agents
What is a Simple Reflex Agent
Uses only built-in knowledge in the form of rules that select action only based on the current percept
- Does not know about performance measure
- No memory and ignores all past percepts
- Typically very fast
What is a Model-based Reflex Agent
Maintains a state variable to keep track of aspects of the environment that cannot be currently observed
- Has memory and knows how the environment reacts to actions
- State is updated using the percept
What is a state and its representation?
States help keep keep track of the environment and the representation can be atomic (just labeled) or factored (vector of attribute values)
What is a Goal-based Agent?
There is a task to reach a defined goal state and then is finished; agent needs to move towards the goal
- Can use search algorithms to plan actions that lead to the goal
- Performance measure is typically the cost to reach the goal
What is a Utility-based Agent
Uses a utility function to evaluate the desirability of each possible states (reward of being in state R(s))
- Choose actions to stay in desirable states
- Performance measure is the discounted sum of expected utility over time
