Download Chapter 11: Artificial Intelligence

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

History of artificial intelligence wikipedia , lookup

Philosophy of artificial intelligence wikipedia , lookup

Transcript
Chapter 11:
Artificial Intelligence
Computer Science: An Overview
Eleventh Edition
by
J. Glenn Brookshear
Copyright © 2012 Pearson Education, Inc. Publishing as Pearson Addison-Wesley
Chapter 11: Artificial Intelligence
• 11.1 Intelligence and Machines
• 11.3 Reasoning
Copyright © 2012 Pearson Education, Inc. Publishing as Pearson Addison-Wesley
11-2
Intelligent Agents
• Artificial Intelligence: the field of computer
science that seeks to build autonomous
machines
• Agent: A “device” that responds to stimuli from
its environment
– E.g., robots, autonomous airplanes, network programs
– Sensors: receiving data from their environments
– Actuators: responding to or affecting their
environments
• The goal of artificial intelligence is to build agents
that behave intelligently
Copyright © 2012 Pearson Education, Inc. Publishing as Pearson Addison-Wesley
11-3
Levels of Intelligent Behavior
• Reflex: actions are predetermined
responses to the input data
• Intelligent response: actions affected by
knowledge of the environment
– E.g., the process of throwing a baseball
• Goal seeking
– E.g., winning a game of chess
– Deliberately forming a plan of action or
selecting the best action
Copyright © 2012 Pearson Education, Inc. Publishing as Pearson Addison-Wesley
11-4
Levels of Intelligent Behavior (Cont.)
• Learning: an agent’s responses improve
over time as the agent learns
– Developing procedural knowledge: involves a
trial and error process
– Storing declarative knowledge: expands or
alters the facts in an agent’s store of
knowledge
Copyright © 2012 Pearson Education, Inc. Publishing as Pearson Addison-Wesley
11-5
Turing Test
• Proposed by Alan Turing in 1950
• Benchmark for progress in artificial
intelligence
• Test setup: Human interrogator
communicates with test subject by
typewriter, without being told whether the
test subject was a human or a machine.
• Test: Can the human interrogator
distinguish whether the test subject is
human or machine?
Copyright © 2012 Pearson Education, Inc. Publishing as Pearson Addison-Wesley
11-6
Turing Test (Cont.)
• Turing’s conjecture: by the year 2000,
machines would have a 30 percent chance
of passing a five-minute Turing test.
• A well-known example of a Turing test
scenario: DOCTOR developed in the mid1960s that simulate psychological interviews
– Computer as analyst, user as patient
– According to some well-defined rules
– “I am tired today”, “Why do you think you’re tired
today?”, “Go on”, “That’s very interesting”…
Copyright © 2012 Pearson Education, Inc. Publishing as Pearson Addison-Wesley
11-7
Figure 10.1 The eight-puzzle in its
solved configuration
Copyright © 2012 Pearson Education, Inc. Publishing as Pearson Addison-Wesley
11-8
Figure 10.2 Our puzzle-solving
machine
• Be able to perceive:
extract the current
puzzle state from
its camera
• Develop and
implement a plan
for obtaining a goal
Copyright © 2012 Pearson Education, Inc. Publishing as Pearson Addison-Wesley
11-9
Components of a Production
Systems
Formulate AI tasks of agents in the context of a
production system:
1. A collection of states: situations that may occur in
the application environment
– Start (or initial) state
– Goal state (or states)
2. A collection of productions: operations that can
be performed to move from one state to another
– Each production may have preconditions
3. A control system: consists of logic that decides
which production to apply next
Copyright © 2012 Pearson Education, Inc. Publishing as Pearson Addison-Wesley
11-10
Reasoning in terms of production
systems
• State Graph: All states and productions
– Nodes: states
– Arrows: productions
• The problem faced by the control system:
finding a sequence of arrows that leads
from the start state to the goal state
– Finding (searching for) a path through a state
graph
Copyright © 2012 Pearson Education, Inc. Publishing as Pearson Addison-Wesley
11-11
Figure 10.4 A small portion of the
eight-puzzle’s state graph
Copyright © 2012 Pearson Education, Inc. Publishing as Pearson Addison-Wesley
11-12
Figure 10.5 Deductive reasoning in the
context of a production system
• The problem of drawing logical conclusions
from given facts, e.g., medical expert systems
– Productions, called inference rules, that allow new
statements to be formed from old statements
– States: collections of statements known to be true at
particular points in the deduction process
– Start state: collection of basic statements (axioms)
– Goal state: any collection of statements that contain
the proposed conclusion
Copyright © 2012 Pearson Education, Inc. Publishing as Pearson Addison-Wesley
11-13
Figure 10.5 Deductive reasoning in the
context of a production system
Copyright © 2012 Pearson Education, Inc. Publishing as Pearson Addison-Wesley
11-14
Reasoning by Search Trees
• A control system’s job involves searching
the state graph to find a path from the start
node to a goal.
• A simple method of performing this search
is to construct a search tree.
• Search Tree: A record of state transitions
explored while searching for a goal state
Copyright © 2012 Pearson Education, Inc. Publishing as Pearson Addison-Wesley
11-15
Figure 10.6 An unsolved
eight-puzzle
Copyright © 2012 Pearson Education, Inc. Publishing as Pearson Addison-Wesley
11-16
Figure 10.7 A sample search tree
Copyright © 2012 Pearson Education, Inc. Publishing as Pearson Addison-Wesley
11-17
Figure 10.8 Productions stacked for
later execution
Copyright © 2012 Pearson Education, Inc. Publishing as Pearson Addison-Wesley
11-18
Heuristic Strategies
• The search tree generated in an attempt to
solve a more complex problem could grow
much larger.
– Developing a full search tree will be impractical
• One strategy: change the order in which
the search tree is constructed
– Breadth-first: layer by layer
– Depth-first: building the more promising paths
rather than horizontal layers. (following intuition)
Copyright © 2012 Pearson Education, Inc. Publishing as Pearson Addison-Wesley
11-19
Heuristic Strategies (Cont.)
• Intuitive methods: need a way of identifying
which of several states appears to be the
most promising
• Heuristic: A quantitative value associated
with each state that estimates the distance
to a goal
• Requirements for good heuristics
– Must be much easier to compute than a
complete solution
– Must provide a reasonable estimate of
proximity to a goal
Copyright © 2012 Pearson Education, Inc. Publishing as Pearson Addison-Wesley
11-20
Heuristic Strategies (Cont.)
• Possible heuristics for the eight-puzzle
– The number of tiles that are out of place
– The number of tiles that are out of place + how
far out of position
• measure the distance each tile is from its
destination and add these values to obtain a single
quantity
Copyright © 2012 Pearson Education, Inc. Publishing as Pearson Addison-Wesley
11-21
Figure 10.9 An unsolved
eight-puzzle
Copyright © 2012 Pearson Education, Inc. Publishing as Pearson Addison-Wesley
11-22
Figure 10.10 An algorithm for a
control system using heuristics
Copyright © 2012 Pearson Education, Inc. Publishing as Pearson Addison-Wesley
11-23
Figure 10.11 The beginnings of our
heuristic search
Copyright © 2012 Pearson Education, Inc. Publishing as Pearson Addison-Wesley
11-24
Figure 10.12 The search tree after
two passes
Copyright © 2012 Pearson Education, Inc. Publishing as Pearson Addison-Wesley
11-25
Figure 10.13 The search tree after
three passes
Copyright © 2012 Pearson Education, Inc. Publishing as Pearson Addison-Wesley
11-26
Figure 10.14
The complete
search tree
formed by our
heuristic
system
Copyright © 2012 Pearson Education, Inc. Publishing as Pearson Addison-Wesley
11-27