* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download John McCarthy defines artificial intelligence as
Survey
Document related concepts
Collaborative information seeking wikipedia , lookup
Incomplete Nature wikipedia , lookup
Agent (The Matrix) wikipedia , lookup
Wizard of Oz experiment wikipedia , lookup
Agent-based model wikipedia , lookup
Ecological interface design wikipedia , lookup
Intelligence explosion wikipedia , lookup
Human–computer interaction wikipedia , lookup
Philosophy of artificial intelligence wikipedia , lookup
Existential risk from artificial general intelligence wikipedia , lookup
Embodied cognitive science wikipedia , lookup
Ethics of artificial intelligence wikipedia , lookup
Transcript
1. How did John McCarthy, the guy who is credited with coining the term, define the term artificial intelligence? John McCarthy defines artificial intelligence as “the goal to develop machines that behave as though they were intelligent.” 2. How did Elaine Rich define the term artificial intelligence? Elaine Rich defines artificial intelligence as “the study of how to make computers do things at which, at the moment, people are better.” This definition sums up what researchers have been doing for the past 50 years and for the future as well. 3. Go to cleverbot.com and interact with the chatterbot for two or three minutes. Copy and paste the most interesting 10 lines (5 of yours; 5 of its) of interaction. User: Are you in love? Cleverbot: Yes I am. User: Who are you in love with? Cleverbot: I love my programmer. User: What is love? Cleverbot: What's love got to do with it? User: What's love but a second hand emotion? Cleverbot: What's emotion but a firsthand account on how a said person sees the world? User: What world is a said person seeing? Cleverbot: Its robot. Bot is what is short for. 4. Say something, in two sentences or less, about Kurt Godel that is at least bit interesting from an AI perspective. Kurt Godel’s completeness theorem and incompleteness theorem are interesting from an AI perspective. The completeness theorem “states that first-order predicate logic is complete” and the incompleteness theorem “shows that there are true statements that are unprovable” in higher-order logics which shows that there are limits of formal systems. 5. Say something, in two sentences or less, about Alan Turing that is at least a little bit interesting from an AI perspective. Alan Turing shows that no program “can decide whether a given arbitrary program will run in an infinite loop which also shows a limit for intelligent programs. 6. Say something, in two sentences or less, about Alonzo Church that is at least a little bit interesting from an AI perspective. Alonzo Church was a mathematician who contributed to number, algorithm, and computability theories which lead to Church’s theorem that “states that only recursive functions can be calculated mechanically and implies that arithmetic procedures cannot be used to decide the consistency of statements formulated in accordance with the laws of arithmetic (Alonzo Church).” 7. Say something, in two sentences or less, about Warren McCulloch and Walter Pitts that is at least a little bit interesting from an AI perspective. Warren McCulloch and Walter Pitts were two people who designed the first mathematical models of neural networks. This is interesting from an AI perspective because neural networks model biological neural networks that process information that can lead to computation of some sort. 8. On top of page 9 your author writes: The above systems offered by AI today are not a universal recipe, but a workshop with a manageable number of tools for very different tasks. Most of these tools are well-developed and are available as finished software libraries, often with convenient user interfaces. The selection of the right tool and its sensible use in each individual case is left to the AI developer or knowledge engineer. Like any other artisanship, this requires a solid education, which this book is meant to promote. What are the “systems” that your author is referring to? Identify at least a few of them. A few of the “systems” that the author are referring to I believe are hybrid systems which are the combination of logic and neural networks, decision tree learning systems which “can represent propositional logic concepts and then be used as expert systems”, and data mining systems which are able to get information from large databases. 9. Does your author suggest that AI is an interdisciplinary pursuit? Rather than merely answering “yes” or “no”, please write a sentence or two that says something meaningful about the proposition. Yes, the author suggests that AI is an interdisciplinary pursuit that is rarely simple and integrates discoveries from logic, operation research, statistics, control engineering, image processing, linguistics, philosophy, psychology, neurobiology, and also areas of specific application. 10. Your author talks a bit about intelligent agents. Google the term, and find three sites that say something interesting about intelligent agents from an AI perspective. Strive to find three sites that are somewhat different in nature. For each site: (1) provide the URL, and (2) write two or three sentence characterization of the site. a. http://www.cs.cmu.edu/~softagents/ This website is Carnegie Mellon University’s Intelligent Software Agents Lab. The lab has developed a multi-agent system infrastructure called RETSINA which consists of interface agents, task agents, information agents, and middle agents. The website includes information on their research, publications, and applications. b. http://aitopics.net/Agents This website has contents dealing with agents that include definitions, readings, and subtopics. The subtopics content include links to multi-agent systems, websearching agents and the semantic web, and social media. This website is laid out nicely and contains many interesting links to other pages with agent information. c. http://www.aosgrp.com/index.html#.UFdKrbKPX7I This website offers software products that “provide the decision-making capability that underlies autonomy.” The website features products, technical services, applications, featured research, and news. I find this website interesting because it talks about how its software is used in many ways. 11. Pick one of the “milestones” listed in Table 1.1. write a 300-500 word passage in your own words on the milestone, and be sure to indicate exactly where you got the material for your passage. Watson is an artificial intelligence, question and answering computer system which is able to understand natural language. Development of Watson began in 2006 and is ongoing to this day. Jeopardy! was picked to test Watson’s ability due to the fact that the game show takes many cognitive abilities that had not been seen to be capable of computers in the past. The cognitive abilities include telling the differences between double meanings of words, puns, rhymes, and inferred hints, rapid responses to questions, and the ability to process large amounts of information that stems from a lifetime of human interactions and cultural aspects of human life. These abilities were goals for the Watson research team to replicate; they focused on three main capabilities which were natural language processing, hypothesis generation, and evidence-based learning (The Science Behind Watson). IBMs DeepQA technology was used for building the above capabilities for Watson. This technology can “perform thousands of simultaneous tasks in seconds to provide precise answers to questions (Watson - A System Designed for Answers, 2011).” The DeepQA technology uses combined techniques which improve accuracy, speed, and confidence. Watson has approximately 200 million pages of natural language content in its “memory.” It was not hooked up to the internet or any other external source while playing Jeopardy!. Watson hardware harnesses POWER7 to be able to execute all the DeepQA task simultaneously; Watson has “90 clustered IBM Power 750 servers which features 32 POWER7 cores that run at 3.55 GHz (Watson - A System Designed for Answers, 2011)”, it runs on the Linux operating system that has 16 terabytes of memory and can operate trillions per second (Watson - A System Designed for Answers, 2011). Watson received the clues on Jeopardy! as texts the same time that human contestants saw the clue, Watson would then parse the clues into different keywords or sentences to find related clues, it would then determine its level of confidence to an answer. If Watson found more algorithms the same then its confidence would increase and so would its chances of being correct (IBM Watson: The Face of Watson, 2011). Although Watson beat both the human players Brad Rutter and Ken Jennings, it does have disadvantages which include understanding contexts of clues and also had a difficult time in understanding short clues. Watson has advantages over human players which mainly have to do with reaction time; human players are signaled by a light that takes tenths of seconds to perceive and Watson was signaled by an electronic signal that only took it 8 milliseconds to activate its buzzer (Thompson, 2010). References Alonzo Church. (n.d.). Retrieved from Encyclopedia Britannica: http://www.britannica.com/EBchecked/topic/117106/Alonzo-Church IBM Watson: The Face of Watson. (2011, February 10). Retrieved from YouTube: https://www.youtube.com/watch?v=WIKM732oEek The Science Behind Watson. (n.d.). Retrieved from IBM: http://www03.ibm.com/innovation/us/watson/science-behind_watson.shtml Thompson, C. (2010, June 16). Smarter Than You Think: What Is IBMs Watson? Retrieved from The New York Times: https://www.nytimes.com/2010/06/20/magazine/20Computert.html Watson - A System Designed for Answers. (2011). IBM Sytems and Technology, 1-5.