Bits and Bytes.docx

1、Bits and BytesBits & Bytes: Lesson 4Computer IQMaking Computers SmartDo you recall this conversation from the film, 2001: A Space Odyssey?Hal, the spacecraft computer, killed a crew member by throwing him into space when the crewman went outside to replace the computer unit. Dave, the hero of the fi

2、lm, has just discovered that Hall has also killed the rest of the crew as they slept.(Dave) Open the pod bay doors, Hal.Hello Hal, do you read me?Do you read me, Hal?Hello Hal, do you read me? Open the pod bay doors, Hal.(Hal) Im sorry, Dave. I cant do that, Dave .the mission is too important for me

3、 to allow you to jeopardize it.(Dave) Ok Hal,Ill go in through the emergency airlock.(Hal) Just what do you think your doing, Dave?I know everything hasnt been quite right with me, but I can assure you its going to be alright again. I feel much better now, I really do.(Hal) Look Dave, I can see you

4、are really upset about this. I know Ive made some very poor decisions lately.I want to help you.Stop, Dave will you stop, Dave? Im afraid, Dave. This film, plus many before and after, fueled our love-fear relationship with smart machines. While machines like Hal, capable of independent thought and e

5、motions, are science fiction, artificial intelligence is currently combining computer science and engineering to create machines that exhibit intelligent behavior, learn useful skills, and adapt to new situations. These machines are embedded within many everyday appliances and tools in our homes, ho

6、spitals, factories, businesses, government agencies, and research facilities. Whole new fields of science and business domains are developing as new applications are invented for the smart computers behind these indispensible machines. And its all based upon some rather basic programming concepts wh

7、ich you will come to realize in this lesson, are very human. Expert system- Computer programs that answer questions and draw conclusions by applying statistical rules and logic.AI and Expert SystemsConventional Artificial Intelligence (AI) is the science of creating computers that mimic what humans

8、refer to as intelligencethe ability to process information and respond appropriately. AI is supported by several types of machine learning technologies including expert systems which apply statistical and reasoning capabilities to huge amounts of data in order to draw conclusions and solve problems.

9、The goal of expert systems is to bring the knowledge of experts to the hands of decision makers who need accurate answers quickly. Computer expert systems are sometimes called knowledge based systems and have been in use since the 1960s. They are created by first gathering the knowledge of human exp

10、erts on some specific, narrow field of knowledge. Then, a knowledge engineer organizes the information with rules that help to focus the knowledge toward solving specific problems.The childhood guessing game in which one child thinks of an animal, and the other child tries to guess it by asking only

11、 yes-or-no questions, simulates an expert system. A process of elimination based upon rules guides a series of questions until the correct animal name is guessed. It might go something like this: “Does it have fur?” “No”“Does it have feathers?” “Yes” “Can it fly?” “No” “Is it black and white?” “Yes”

12、 “Is it a penguin?” “Yes”Fuzzy Logic- A system that relies upon “mostly true” and “mostly false” conditions to arrive at more realistic “most likely” conclusions.Expert systems contain information supplied by experts on a single topic and usually involve a “dialog” with the user. This problem-solvin

13、g system relies upon rules that guide the dialog toward the discovery of the solution that fits all of the responses given by the user. A recent addition to expert systems is the inclusion of “fuzzy logic” which doesnt rely on definite yes-or-no answers, but rather on conditions of “mostly true” or

14、“mostly false.” These less definite answers result in more realistic “most likely” conclusions.Expert systems are employed quite commonly in: Wizards that help you install software Artificial intelligence of computer games Diagnostic programs for everything from identifying skin diseases to getting

15、help with your computer system Decision making in business, science, engineering, and the militaryComputational Intelligence and Artificial Neural Networks Neural network- A collection of interconnected processing units which resemble brain neurons and are capable of complex analysis of data.A speci

16、alized area of artificial intelligence called computational intelligence includes the application of artificial neural networks. When groups of processing unitssometimes referred to as nodesare interconnected in a fashion resembling the neurons in the brain, an artificial neural network can be const

17、ructed. An artificial neural network “learns” by observing complex relationships between inputs and outputs and inferring from observations. This is particularly useful in applications where the complexity of the data or the difficulty of the task makes traditional program design extremely complicat

18、ed or impractical. Artificial neural networks are especially useful in tasks such as: Pattern recognition in military radar systems Face identification in Homeland Security operations Sequence recognition in gesture, speech, writing Medical diagnosis with EMR technology Detection of abnormal charges

19、 on credit cards Data mining or knowledge discovery in databases Identification of e-mail spam Robotics and Artificial IntelligenceTodays robots are simply machines with enough AI to do useful tasks. And the tasks run the gamut from entertainment in the form of Aibo, and the robotic pet, to Roomba,

20、the robotic vacuum cleaner, to Penelope, the surgical assistant. Robots come in all shapes and sizes. Depending upon how you define the word “robot,” you could be talking about the “smart wiper” on your car windshield that is “just a robotic arm” or a machine the size of a vehicle such as the Mars R

21、over or the trucks in the DARPA Challenge road race.It appears that robots will definitely be part of our future of smart machines. Analysts estimate that about 4 million Roomba vacuum robots will be in homes by 2007 and some people in Japan foresee 39 million household robots in use by the end of t

22、he decade. Add to that the 1.1 million industrial robots in the U.S. and you have a robot population greater than the population of California or Spain! Research and development in AI and robots includes such projects as: Systems for surgical assistance Solar-powered underwater vehicles Robotic assi

23、stants for the elderly Space exploration rovers Bomb-sniffing military robotics Childcare workersIDF American Andros Explosive Ordinance Device (EOD) robotSince the earliest days of computers and AI, great debate has raged over a definition of intelligence and creativity, whether or not a machine ca

24、n actually understand and be intelligent, and even, if the distinction between human and machine will eventually disappear. Various schemes to distinguish between computer and human intelligence and creativity have appeared over the past several decades. Perhaps you have heard of the Turing Test in

25、which a judge asks questions to an entity in another room. If at some point the judge cannot tell if the answers are coming from a person or a computer, the computer is declared intelligent. And of course, weve all heard of the famous computer versus human chess games. While philosophers will contin

26、ue to explore the distinctions between humans and machines, computer scientists will forge ahead with research and developments of more and more amazing machines. Ultimately, computer IQ comes down to machines following the directions given to it by programmers and engineers.Get Real1. Explore an ex

27、pert system by visiting The Whale Watcher at http:/www.aiinc.ca/demos/whale.html. Does this system employ strict true/false logic or fuzzy logic?2. Review the basic characteristics of expert systems, fuzzy logic, neural networks, and robots. Match the tasks on the left with the technology on the lef

28、t that would best be able to accomplish the task described.Task to accomplishTechnology _ Discover the most likely cause of a panic attackA. Expert systems_ Identify wild flowers seen on a hikeB. Fuzzy logic_ Inspect mines for dangerous gasesC. Neural networks_ Uncover forged artworkD. Robotics Here

29、 are some possible answers:1. After identifying several hypothetical sighting of whales, it appears that this system employs strict true/false logic. All of the response choices are yes/no or either/or.2. Task to accomplishTechnology _B_ Discover the most likely cause of a panic attackA. Expert syst

30、ems A_ Identify wild flowers seen on a hikeB. Fuzzy logic_D_ Inspect mines for dangerous gasesC. Neural networks_C_ Uncover forged artworkD. Robotics Teaching Computers to Make DecisionsComputers are “taught” to display intelligence by making decisions through various coding techniques. Collectively

31、 these techniques are called “selection structures”; the lines of code instruct the computer to “select” between various options based upon particular conditions. If-StatementsIf-statement- A computer instruction that decides upon an action based upon the evaluation of a condition. The most basic se

32、lection structure is the if-statement. In the real-world we process if-statements so often and with such speed that we barely give it any thought at all. You have likely experienced these opportunities to process an “if” decision:If its raining outside, you grab an umbrella.If steak is on sale, you buy it.If the traffic signal is red, you stop.If you see a friend, you shout, “Hi!”In all of these