AI Quiz 1
arise when the agent has no idea about the states and actions of its environment.
Exploration problems
The process in a one dimensional state-space landscape, where elevation corresponds to an objective function, and the aim is to find the highest peak -a global maximum, is called
Hill climbing
Knowing something about the current state of the environment is not enough information to decide what to do. An agent needs some sort of ________________________ that describes situations that are desirable
goal information
Informed search methods
have access to a heuristic function h(n) that estimates the cost of a solution from n.
Percept Sequence
is the complete history of everything the agent has ever perceived
What type of search does this describe ?
local beam
If there is a one to one mapping between concepts and memory locations, we call that a
localist representation
Select all of the statements that are true about Atomic Representation:
long answers
A __________________ path cost always increases as you go along a path, because action costs are always positive
monotonic
A key property of A * search is/are
never overestimates the costs to reach a goal
Select all of the applicable characteristics of the Two Player Zero Sum game
no win-win outcome, perfect information, deterministic
Uninformed search methods have access only to the
problem definition
allows us to make a decision without examining every state
pruning
Beam search
puts a limit on the size of the frontier; that makes it incomplete and suboptimal, but it often finds reasonably good solutions and runs faster than complete searches
Stochastic games include a ____________________ such as throwing dice
random element
Deep learning
refers to machine learning using multiple layers of simple adjustable computing elements.
ProbCut is a version of Alpha-beta that uses _______________ to lessen the change that the best move will be pruned
statistics
games include CHANCE nodes
stochastic
What are the common player names for the Two-Player zero sum game ?
Max and min
The recursive algorithm that proceeds all the way down to the leaves of the tree and then backs up the minimax values through the tree as the recursion unwinds is the
MiniMax Search
In a minimax tree with a branching factor of 3 and depth 2 (one max layer, one min layer with 3 nodes, and a leaf layer with 9 total nodes) what is the maximum number of nodes that can be pruned by alpha-beta pruning?
4 ( 2 each in the right most and 2 leaf layers)
What is one of the limitations of Game Search ?
All algorithms must make assumptions and approximations
pruning prunes branches that have no effect on the final evaluation
Alpha-Beta
In a full-depth minimax search of a tree with depth D and branching factor B, with α−β pruning, what is the minimum number of leaves that must be explored to compute the best move?
BD/2
Uniform-cost search
expands the node with lowest path cost, g(n), and is optimal for general action costs
Knowledge based agents can accept new tasks in the form of
explicitly described goals
algorithms operate by searching from a start state to neighboring states, without keeping track of the paths OR the states that have been reached
Local search
Which of the following statements about alpha-beta pruning are true or false?
None are true. Alpha-beta always finds the optimal strategy against an opponent that plays optimally. If an ordering heuristic is available we can expand nodes in an order that maximizes pruning. Alpha-beta requires less run time than minimax except in contrived cases.
a purely random walk that moves to a successor state without concern for the value i.e. picks a random move
Simulated annealing search
The standard names for the operations to add new sentences to a knowledge base and query what is known in the knowledge base are
Tell and Ask
Q: A∗ is of no use in robotics because percepts, states, and actions are continuous.
false A* search is often used in robotics. the space can be discretized or skeletonized
Q: Depth-first search always expands at least as many nodes as A* search with an admissible heuristic.
false. A lucky depth first search might expand exactly d nodes to reach the goal. A* largely dominates any graph search algorithm that is guaranteed to find optimal solutions.
pruning prunes moves that appear to be poor, but might be good
forward
Q: h(n) = 0 is an admissible heuristic for the 8-puzzle.
True h(n)=0 is always an admissible heuristic since costs are nonnegative
Q: Breadth-first search is complete even if zero step costs are allowed.
True: depth of the solutions matters for depth first search not cost
What does the Best-First-Search function return ?
a solution node or a failure
job of AI is to design an ___________________ that implements the agent function, i.e., the mapping from percepts to actions
agent program
Monte Carlo Tree Search estimates the value of a state as the ____________ over a number of simulations of complete games
average utility
Agent programs combine particular components in particular ways to generate actions.
basic kind of agent program that embodies the principles underlying almost all intelligent systems: Simple reflex agents, model based, goal based
A useful way to visualize a search is to ______________ in the state space
draw contours
atomic representation
each state of the world is indivisible, i.e., it has not internal structure
The local search algorithm that is explicitly motivated by the metaphor of natural selection in biology is known as the
evolutionary algorithm
performance measure
the sequence is desirable, then the agent has performed well
The goal of learning is to minimize the _____________ of problem solving, trading off computational expense and path cost
total cost
situation that an agent can find itself in
world state, state, representational state, initial