CS1000 Introduction to Computer Science by fanzhongqing


									        Chapter 10: Artificial Intelligence

   Outline
       Introduction                   Social Issues

       Main tasks                     Applications

       Knowledge
                                      Virtual Machine

       Recognition tasks
                                  Algorithmic Foundations
       Reasoning tasks

   Artificial Intelligence (AI)
       The part of computer science that exploits human
        intelligence in deriving computer algorithms
       The more we know how human intelligence works, the more
        “intelligent” computer-based solutions can be achieved.
       Is the machine “intelligence” not enough?
            No …
            Because we still have a multitude of problems that cannot be
             solved by a computer at all.
            Because we still have a multitude of (unfortunately practice-
             close) problems that are solvable but need years or hundreds
             of years to come to a result.
            Because humans are “more intelligent” than computers in a
             variety of situations.

   Turing Test
       A. Turing proposed in 1950s a method to test the
        intelligence of machines
       Human interrogates two entities
            Entity 1: Human
            Entity 2: Computer
       Interrogator is not allowed to see where an answer comes
        from (e.g. answers are printed prior to inspection)
       Test:
            If as a result of the questioning, the interrogator is not able to
             determine which answer originates from the computer and
             which answer originates from the human, then the computer
             has passed the Turing intelligence test.

   Examples:
       Is your partner in a chess game a computer or a human?
       Are you communicating (e.g. by email) with a computer or
   Problems with Turing Test:
       Measures intelligence of computers based on human
       This kind of comparisons are no more of interest, since
        computer should support humans and NOT replace them.
       Also:
            What about interrogator’s intelligence?
            What types of questions are representative ones?
            …

   Roughly speaking there were two phases in the
    history of AI
       Initial euphoric phase with great expectations, which were
        not realized. (science fiction remained science fiction!)
       Current phase with more realistic expectations and with
        valuable results.
   Observe:
       Artificial: may mean apparent, non-genuine, mimed, …
       Intelligence: may mean thought-based information, the right
        information in the right place/time, …

Main Tasks
   Let us focus on three human tasks:
       Computational tasks
            Adding columns of numbers
            Sorting a list of numbers
            Search a given name in telephone book
            Manage the payroll of a company
            …
       Recognition tasks
            Recognizing your best friend
            Understanding the spoken word
            Finding the tennis ball in the grass in your backyard
       Reasoning tasks
            Planning what to wear today
            Deciding on the strategic directions of a company in the next five years
            Running an alarm after an earthquake

Main Tasks
   Humans and computational tasks:
        Humans can follow algorithmic steps in order to come to a result
        But results should be found very quickly:
             Humans make mistakes
             Get bored
             And become sloppy
   Computers and computational tasks
        Computers are the specialists in this domain (and only this
        They don’t get bored and don’t become sloppy
        They are very fast in following stepwise instructions
        This is why we emphasized the step-by-step instruction processing
         so much in the early chapters
    computers are better than humans in performing
    computation tasks provided that an efficient step-by-step
    solution (algorithm) is used                                        7
Main Tasks
   Humans and recognition tasks
       We, humans, are good in recognition tasks
       We exploit our sensory-recognition-motor skills very
       We receive information through our senses (hearing, seeing)
       We can recognize the information we “sensed”
       And we usually response to the received information by
        “doing” something (e.g. movement)
       Compare: an infant (a few weeks old) is able to recognize
        his/her mom’s face!!, BUT the same person will need in
        general at least six/seven years in order to perform basic
        arithmetic operations

Main Tasks
       How do we recognize things?
            This topic is shared by different science disciplines
            Consider recognizing your best friend:
                  You have a “database” of pictures in your brain
                  When you see your best friend, you match his/her picture against
                   your “file”
                  If the match was successful you will probably laugh and select a
                   specific manner of communication (e.g. language)
                  Moreover:
                       You can recognize your best friend’s sister even if you have

                         never seen her before
                       Thus you don’t need exact or complete information for


   Computers are not that good in recognition tasks
Main Tasks
   Humans and reasoning tasks
       We use also here a large storehouse of information (e.g.
       This information consists not only of immediate facts like images
        but also of cause-and-effect rules
       Example: wearing a coat in a winter day
            You decide to wear a coat because you know by experience that you
             would be uncomfortable without a coat
            Reasoning steps
                I don’t want to be cold

                If it is winter and I don’t wear a coat, then I will be cold

                It is winter now

                Conclusion: I will wear a coat

       To mimic reasoning steps using a computer is a challenging task
       We, humans, can often come to a conclusion even if we do not
        have enough information, we may be ambiguous, and exploit
        intuition                                                     10
Knowledge Representation
   From the previous assessment it follows that AI tries
    to improve computer ability for solving recognition
    and reasoning problems
   Likewise, it was mentioned that for these tasks to be
    achieved, information bases (e.g. picture databases)
    are needed
   Thus: we need to tackle the problem of representing
    information (knowledge) in a computer system
   Knowledge:
       Facts or truths about some topic
       Rules for gaining new facts

Knowledge Representation
   How to represent knowledge?
       Natural language:
            For example:
                   “Spot is a brown dog, and, like any dog, has four legs and a tail. Also, like
                   any dog, Spot is a mammal, which means Spot is warm-blooded.”
            Notice that these “strings” have a meaning and have to be treated as
       Formal language:
            x: entity e.g. a dog
            A(x): x has attribute A
            Above example:
                  Dog(Spot)
                  Brown(Spot)
                  For all x: Dog(x)  FourLegs(x)
                  For all x: Dog(x)  hasTail(x)
                  For all x: Dog(x)  isMammal(x)
                  For all x: isMammal(x)  isWarmBlooeded(x)

Knowledge Representation
    Pictorial representation:
         Give the picture of Spot showing that it is brown and has four
          legs and a tail
         +: additional information can be contained
         -: attributes warm-blooded and mammal cannot be
          represented, additional text is necessary
    Graphical representation
         Semantic net
                Mammal                Warm blood

             is a
                                      has           4 legs
                     Dog                    has
                           is color
                    Spot                    brown
Knowledge Representation
   What type of representation to use:
       It depends on the application domain
       E.g. for computer vision, rather pictorial
       Semantic nets are extensible and therefore are appropriate
        for capturing non-complete knowledge
       Natural languages are not as exact as formal languages.
   General criteria for representation methods:
       Adequateness
       Efficiency
       Extensibility
       Appropriateness

Recognition Tasks
   AI sometimes tries to mimic the way of human
    thinking ( brain functions)
   But how does our brain function?
       1011-1012 neurons
       Stimuli “enter” a neuron though dendrites
       Stimuli “exit” a neuron through axons
       Axons connect to other dendrites by synapses
       A neuron gathers stimuli from dendrites
       If the sum of signals is higher than a threshold, the neuron
        fires; it sends a new signal down its axon and affects other
        neurons in its neighborhood

Recognition Tasks
   Neural Network
       Computer scientists have developed (artificial) neural networks to
        simulate the work of our brain in order to solve problems related to
       Hardware or software implementation
       Each node represents the nucleus of a neuron
       Each node has a specific threshold value
       Each node has a number of weighted input lines; the dendrites
       Each node has a number of output lines; the axons/synapses
        If sum of weighted inputs >= threshold, then output is activated

                   Weight 1
                                  Node      Weight
                   Weight 2     (Nucleus)

                   Weight 3                                            16
Recognition Tasks
   Example: recognizing two character patterns:
       Node: Class A
            Input: all shapes of an “A”
            Output: to “Different” and “Equivalent” nodes
       Node: Class B
            Input: all shapes of a “B”
            Output: to “Different” and “Equivalent” nodes
       Examples:
            Inputs: A and A  Equivalent is activated
            Inputs: B and B  Equivalent is activated
            Inputs: A and B  Different is activated
   How to come to the correct weights:  training
       Training phase (done automatically by a “trainer” module):
            Network is fed by an initial set of input with known output
            Weights are iteratively adjusted until actual output is close enough to
             desired output
            Compare: training a dog until it “hardwires” things in its brain     17
Recognition Tasks
   Example network
        A       2
                2   Class A       2     Different
                       1                    3
        A                         -2

    B           2

            2       Class B            Equivalent
    B                  1          -2       -3
Recognition Tasks
   Neural networks have been applied in a wide range
    of areas:
       Handwriting recognition
       Speech recognition
       Recognizing bad credit risks in loans
       Predicting the odds of cancer susceptibility
       Limited visual recognition
       Segmenting magnetic resonance images in medicine
       Adapting mirror shapes for astronomical observations
       Discovering a good routing algorithm in a computer network

    Reasoning Tasks
   How do humans reason in a logical way?
   Example: Triage center in a hospital
       Understanding the situation
            Capability of staff
            Availability of resources
            Older experiences
            …
       Conclusion:
            Patient A: first priority
            Patient B: second priority
            …
   In AI, rule-based systems (or expert systems) are
    used to emulate this kind of reasoning
        Reasoning Tasks
   A rule-based system consists of:
       A knowledge base and
       An inference engine
   Knowledge base:
       Set of facts and rules about a special domain
       Examples (compare Prolog facts and rules):
        F1: Lincoln was president during Civil War
        F2: Kennedy was president before Nixon
        F3: FDR was president before Kennedy
        R1: if X was president before Y, then X precedes Y
        R2: if X was president before Z and Z precedes Y, then X precedes Y

        Reasoning Tasks
   Inference Engine:
       How to generate new facts from the knowledge base
       Example 1:
        F2: Kennedy was president before Nixon
        R1: if X was president before Y, then X precedes Y
        F2 and R1 lead to a new fact F4:
        F4: Kennedy precedes Nixon
       Example 2:
        F3: FDR was president before Kennedy
        R2: if X was president before Z and Z precedes Y, then X precedes Y
        F3, F4, and R2 lead to a new fact F5:
        F5: FDR precedes Nixon
          Reasoning Tasks
   Inference Engine:
       Basic meta-rule we are using to infer new knowledge is:
        Given the fact A and the rule if A then B,
        then infer a new fact B.
       Using this meta-rule, called modus ponens, new facts can be “learned”
        from older ones.
       These new facts may let other rules in the knowledge to be used in
        order to generate further new facts, and so on.
       The process of inference is repeated until no new rules can be inferred.
       Two policies for inference:
        Forward chaining: Match facts to the if-branches of rules (like our
        Backward chaining: Match facts to the then-branches of rules

To top