robotics versus artificial intelligence. search. searchsearch “all ai is search” “all ai is...

19
Robotics Robotics Versus Versus Artificial Artificial Intelligen Intelligen ce. ce. Search Search

Post on 21-Dec-2015

218 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: Robotics Versus Artificial Intelligence. Search. SearchSearch “All AI is search” “All AI is search”  Game theory  Problem spaces Every problem is a

Robotics VersusRobotics Versus

Artificial Artificial Intelligence. Intelligence.

SearchSearch

Page 2: Robotics Versus Artificial Intelligence. Search. SearchSearch “All AI is search” “All AI is search”  Game theory  Problem spaces Every problem is a

SearchSearchSearchSearch ““All AI is search”All AI is search”

Game theoryGame theory Problem spacesProblem spaces

Every problem is a Every problem is a feature spacefeature space of all possible of all possible (successful or unsuccessful) solutions.(successful or unsuccessful) solutions.

The trick is to find an efficient search strategy in The trick is to find an efficient search strategy in this space.this space.

Page 3: Robotics Versus Artificial Intelligence. Search. SearchSearch “All AI is search” “All AI is search”  Game theory  Problem spaces Every problem is a
Page 4: Robotics Versus Artificial Intelligence. Search. SearchSearch “All AI is search” “All AI is search”  Game theory  Problem spaces Every problem is a
Page 5: Robotics Versus Artificial Intelligence. Search. SearchSearch “All AI is search” “All AI is search”  Game theory  Problem spaces Every problem is a

Example of an Intelligent ActionExample of an Intelligent ActionExample of an Intelligent ActionExample of an Intelligent Action

Getting ready to come to classGetting ready to come to class Describe so a machine could do itDescribe so a machine could do it

search among alternatives (car or bus)search among alternatives (car or bus) represent the knowledgerepresent the knowledge

This requires a lot of knowledge…!

Page 6: Robotics Versus Artificial Intelligence. Search. SearchSearch “All AI is search” “All AI is search”  Game theory  Problem spaces Every problem is a

SearchSearchSearchSearch Example: Four three-letter Example: Four three-letter crossword crossword puzzlepuzzle

Search problem is find correct puzzleSearch problem is find correct puzzle Approaches:Approaches:

wordword fill: fill: take word, take word, put it to space, put it to space, if contradiction, backtrackif contradiction, backtrack

spacespace fill: fill: take vertical or horizontal spacetake vertical or horizontal space select a word with this lengthselect a word with this length put it into space,put it into space, if contradiction, backtrackif contradiction, backtrack

Many other strategies: homework, find Many other strategies: homework, find space and operators in it, discuss space and operators in it, discuss backtracking strategybacktracking strategy

Cat dog cam may mom sit mit

C A MA O

T O M

Page 7: Robotics Versus Artificial Intelligence. Search. SearchSearch “All AI is search” “All AI is search”  Game theory  Problem spaces Every problem is a

Blind SearchBlind SearchBlind SearchBlind Search

Search depends only Search depends only on node’s position in on node’s position in the search treethe search tree

Two basic blind Two basic blind searches:searches: depth-firstdepth-first breadth-firstbreadth-first

Problem:Problem: define depth- define depth-first search for the first search for the above problemabove problem

Problem:Problem: define define breadth-first algorithm breadth-first algorithm for the above problem.for the above problem.

Repeat both for each Repeat both for each of the approaches of the approaches outlined above.outlined above.

Called also search strategies

Page 8: Robotics Versus Artificial Intelligence. Search. SearchSearch “All AI is search” “All AI is search”  Game theory  Problem spaces Every problem is a

Depth-First Search Pseudo-CodeDepth-First Search Pseudo-CodeDepth-First Search Pseudo-CodeDepth-First Search Pseudo-Code 1. Set L to list of initial nodes1. Set L to list of initial nodes 22. n. n = head(L), Empty(L) ==> fail = head(L), Empty(L) ==> fail 3. If3. If n n=goal, stop, return it and return the path =goal, stop, return it and return the path

leading to itleading to it 4. pop(L), push(L) all 4. pop(L), push(L) all nn’s children, ’s children, 5. go to5. go to step 2 step 2

Head = car in LISP

empty = null in LISP

push(L)pop(L)

Depth First is based on a stack, L

Page 9: Robotics Versus Artificial Intelligence. Search. SearchSearch “All AI is search” “All AI is search”  Game theory  Problem spaces Every problem is a

Breadth-First SearchBreadth-First SearchBreadth-First SearchBreadth-First Search

1. Set L to list of initial nodes1. Set L to list of initial nodes 2. 2. nn=head(L), Empty(L)==> fail=head(L), Empty(L)==> fail 3. If 3. If nn=goal then stop and return it and path=goal then stop and return it and path 4. Dequeue(L), Enqueue(L) all 4. Dequeue(L), Enqueue(L) all nn’s children’s children 5. 5. Go toGo to step 2 step 2

Depth First is based on a queue, L

Means, remove from queue

Means, add to queue

Page 10: Robotics Versus Artificial Intelligence. Search. SearchSearch “All AI is search” “All AI is search”  Game theory  Problem spaces Every problem is a

Heuristic SearchHeuristic SearchHeuristic SearchHeuristic Search

Meta-level reasoningMeta-level reasoning heuristicheuristic function aids in selecting which part function aids in selecting which part

of search tree to expandof search tree to expand trade-offtrade-off between time to compute heuristic between time to compute heuristic

function and to expand the treefunction and to expand the tree

Page 11: Robotics Versus Artificial Intelligence. Search. SearchSearch “All AI is search” “All AI is search”  Game theory  Problem spaces Every problem is a

Other IssuesOther IssuesOther IssuesOther Issues

BacktrackingBacktracking chronological backtracking as in Prologchronological backtracking as in Prolog dependency-directed backtrackingdependency-directed backtracking

Search direction:Search direction: forward (toward goal)forward (toward goal) backward (from goal)backward (from goal)

and math proving problemsand math proving problems Bi-directionalBi-directional

and building tunnel storyand building tunnel story

Page 12: Robotics Versus Artificial Intelligence. Search. SearchSearch “All AI is search” “All AI is search”  Game theory  Problem spaces Every problem is a

Search Examples ISearch Examples ISearch Examples ISearch Examples I

Game playingGame playing chesschess backgammonbackgammon

Finding path to goalFinding path to goal Missionaries and cannibalsMissionaries and cannibals Towers of HanoiTowers of Hanoi Sliding TileSliding Tile games (15, 8) , puzzles games (15, 8) , puzzles

Page 13: Robotics Versus Artificial Intelligence. Search. SearchSearch “All AI is search” “All AI is search”  Game theory  Problem spaces Every problem is a

Search Examples IISearch Examples IISearch Examples IISearch Examples II

finding a goalfinding a goal cryptoarithmeticcryptoarithmetic n-queensn-queens ““mutilated checkboard” or “tough nut” of mutilated checkboard” or “tough nut” of

McCarthy problem.McCarthy problem.

Page 14: Robotics Versus Artificial Intelligence. Search. SearchSearch “All AI is search” “All AI is search”  Game theory  Problem spaces Every problem is a

Example Applications of search Example Applications of search Example Applications of search Example Applications of search

Expert SystemsExpert Systems Natural Language ProcessingNatural Language Processing VisionVision RoboticsRobotics

Page 15: Robotics Versus Artificial Intelligence. Search. SearchSearch “All AI is search” “All AI is search”  Game theory  Problem spaces Every problem is a

Search: Game TheorySearch: Game TheorySearch: Game TheorySearch: Game Theory

9!+1 = 362,880

Robot interaction with humans, other robots and environment can be described in terms of a game

tic-tac-toe,

Page 16: Robotics Versus Artificial Intelligence. Search. SearchSearch “All AI is search” “All AI is search”  Game theory  Problem spaces Every problem is a

Game playing Game playing Game playing Game playing Programs that:Programs that:

take advantage of the computer's ability to examine a large take advantage of the computer's ability to examine a large number of possible moves in a short period of time and number of possible moves in a short period of time and

logically assess their probable success or failure logically assess their probable success or failure

have been already developedhave been already developed They were used for game playing in:They were used for game playing in:

tic-tac-toe, tic-tac-toe, checkers, checkers, chess, chess, and the Oriental game called Go. and the Oriental game called Go.

Page 17: Robotics Versus Artificial Intelligence. Search. SearchSearch “All AI is search” “All AI is search”  Game theory  Problem spaces Every problem is a

Recent Trends in Artificial IntelligenceRecent Trends in Artificial IntelligenceRecent Trends in Artificial IntelligenceRecent Trends in Artificial Intelligence Intelligent agentsIntelligent agents

Experimental AI software designed to sift through masses of Experimental AI software designed to sift through masses of information made available on the evolving Information information made available on the evolving Information Superhighway of cyberspace to suggest topics of interest or Superhighway of cyberspace to suggest topics of interest or importance for an individual. importance for an individual.

Artificial life Artificial life A field of AI research that studies the adaptive control systems of A field of AI research that studies the adaptive control systems of

insects and other ecological systems and reproduces them in insects and other ecological systems and reproduces them in robotic insectoids.robotic insectoids.

More recently researchers have been working on robots More recently researchers have been working on robots with the intelligence of a two-year-old child.with the intelligence of a two-year-old child.

Intelligent Agents and Artificial Life are now part of Intelligent Agents and Artificial Life are now part of robotics robotics

Page 18: Robotics Versus Artificial Intelligence. Search. SearchSearch “All AI is search” “All AI is search”  Game theory  Problem spaces Every problem is a

Your tasksYour tasksYour tasksYour tasks Search: Search: give an example of search problem that can be solved give an example of search problem that can be solved

using Lisp and that has not been presented so far in the classusing Lisp and that has not been presented so far in the class Blind and informed search: Blind and informed search: for your problem, give an example of for your problem, give an example of

blind search and informed search. Create a powerful heuristic to solve blind search and informed search. Create a powerful heuristic to solve it.it.

Games: Games: examples of games include tic-tac-toe, checkers, examples of games include tic-tac-toe, checkers, chess, go, othello, etc. What kind of game can be a good chess, go, othello, etc. What kind of game can be a good choice for our robot-guard in the FAB building choice for our robot-guard in the FAB building environment?environment?

Expert system: Expert system: what kind of expert knowledge is neede for our what kind of expert knowledge is neede for our walking guard, in addition to knowledge about the FAB building walking guard, in addition to knowledge about the FAB building geometry, distances, what is in which office, structure of ECE geometry, distances, what is in which office, structure of ECE department and its people, etc.?department and its people, etc.? How is this knowledge stored and accessed?How is this knowledge stored and accessed?

Page 19: Robotics Versus Artificial Intelligence. Search. SearchSearch “All AI is search” “All AI is search”  Game theory  Problem spaces Every problem is a

Possible Lisp Problems for the examPossible Lisp Problems for the examPossible Lisp Problems for the examPossible Lisp Problems for the exam 1. Mouse in Labyrinth: 1. Mouse in Labyrinth: design a Lisp algorithm that will be able to go out design a Lisp algorithm that will be able to go out

of every labyrinth. Discuss the program strategy versus some knowledge of the of every labyrinth. Discuss the program strategy versus some knowledge of the labyrinth geometry. What kind of concepts introduced so far in the class may labyrinth geometry. What kind of concepts introduced so far in the class may be useful? Discuss using depth-first, breadth-first search and other search be useful? Discuss using depth-first, breadth-first search and other search strategies.strategies.

2. Obstacle Avoiding robot: 2. Obstacle Avoiding robot: design a Lisp program that will design a Lisp program that will simulate a turtle avoiding obstacles. In contrast to your homework 2, however, simulate a turtle avoiding obstacles. In contrast to your homework 2, however, the turtle can recognize type of the obstacle and select the best strategy. The the turtle can recognize type of the obstacle and select the best strategy. The best strategy is not necessarily to follow the outline of the obstacle, but to go best strategy is not necessarily to follow the outline of the obstacle, but to go straigth between two obstacles, sometimes closer to one of them. This is a straigth between two obstacles, sometimes closer to one of them. This is a useful subproblem of avoiding stationary obstacles in the corridor or room for useful subproblem of avoiding stationary obstacles in the corridor or room for our robot. our robot. A program like this was written by Mike Burns. Read about Voronoir diagrams A program like this was written by Mike Burns. Read about Voronoir diagrams

and their uses in robotics.and their uses in robotics. Think how these diagrams can be improved if you have more knowledge of Think how these diagrams can be improved if you have more knowledge of

obstacles.obstacles.