NOTE: Free essay sample provided on this page should be used for references or sample purposes only. The sample essay is available to anyone, so any direct quoting without mentioning the source will be considered plagiarism by schools, colleges and universities that use plagiarism detection software. To get a completely brand-new, plagiarism-free essay, please use our essay writing service.
One click instant price quote
... if you describe the rust spots on your car, MYCIN will likely diagnose your car with meningitis as well. (Mechner, 1998). At the heart of many game-playing programs is an evaluation function. Early on, game developers quickly encountered the knowledge-acquisition bottleneck and traded quality of knowledge for speed of the program. Simple evaluation functions, linear combinations of easily identifiable features, were the mainstay of computer game programs for many decades. Alternative approaches, such as modeling human cognitive processes, turned out to be much harder to do than initially expected and generally resulted in poor performance.
Game programmers quickly learned that a little heuristic knowledge, when combined with deep search, can produce amazing performance results. Indeed, one could argue that the viability of brute-force search, once a term with negative connotations in the AI community, is one of the main research results from games-related research. The term search, in this context, refers to a simple algorithm known as "minimax, " used in all successful modern chess programs. When choosing from among several moves, the logic goes, explore as many branches of the "game tree" as possible: first try all your currently possible moves, then try all your opponent's possible responses to each of your moves, then try your responses to those responses, and so on, until you run out of time to think. Finally, pick the move that leads to the best possible position, assuming your opponent will offer the strongest possible resistance. (Mechner, 1998). Today, forty-seven years after the minimax algorithm was first proposed as an approach to computer chess, Deep Blue can examine a staggering 200 million positions per second.
That raw power, combined with clever refinements in the algorithm, enable Deep Blue to think fourteen or fifteen moves ahead -- and as deep as forty or forty-five moves when necessary. As a result, the program has beaten Kasparov, arguably the strongest chess player ever. The grand challenge of computer chess has been met. Or has it? (Mechner, 1998). Over the years, AI investigators have gotten used to swallowing big, nasty doses of reality.
Many of them, therefore, are greeting Deep Blue's success as a refreshing change of pace. Still, others are left cold. They deride Deep Blue for using "brute force" -- for having nothing whatever to do with human intelligence. If the purpose of AI is to learn enough about intelligence to create intelligent computer programs, as the pioneering investigator John McCarthy, professor of computer science at Stanford University, suggests, then computer chess has little to do with AI, and Deep Blue is just a triumph of engineering. When human players look ahead, they examine just a few moves at various choice points -- or just one.
Instead of considering billions of positions, they consider dozens. When a reporter asked the former world chess champion Jose Capablanca how many moves ahead he looked while playing, Capablanca replied: "Only one, but it's always the right one. " (Mechner, 1998). Introductory programming curricula are an increasingly common, if somewhat more unusual, place to find robots. There are many who argue that a robot gives a concrete, hands-on, pragmatic way of forcing novice programmers to confront software engineering challenges -- unpredictability, changing requirements, need for maintenance (beyond initial development) of code -- that substantially augment the introductory experience. These perspectives are not universally held, however, and there was discussion of the trade-offs between the potentially enhanced experience and the additional overhead imposed. (Davis, 1999).
Not surprisingly, given the setting, several researchers used robotics in the AI curriculum. These uses range from more traditional robotic applications to integration of robotics into a broad portion of the AI curriculum. However, not all the robotics applications were in intuitively technologically based areas. Other robotics-based courses included freshman writing seminars and introduction s to engineering for humanities majors. We saw a program for Irish elementary schoolchildren who use robots to mount full-scale story-telling exhibitions and a soft and fuzzy robot intended to help elementary school children think about learning. (Takahashi, 2000). "Would you trust your autonomous system if your life depended on it?" When this question was posed to the Robust Autonomy Workshop participants, the response was a bit under-whelming -- 3 of the 40 participants raised their hands. When posed to the entire Spring Symposium community, only one additional person raised her hand.
This lack of trust, even by the developers, provides the setting and the impetus to address the challenges in designing and deploying autonomous systems that are required to continue operating in the presence of failures and unanticipated events. Invited speaker Richard Doyle (Jet Propulsion Laboratory) kicked off the workshop with an inspiring discussion on future National Aeronautics and Space Administration (NASA) missions that will demand robustness -- missions such as a Europa hyd robot that will be required to perform its mission under Europa's "seas" without the luxury of communication with Earth. (Atkin, 1999). The technical presentation topics ranged from lessons learned from deployed systems to tools and architectures for developing robust systems to managing environmental uncertainties. These presentations sparked many lively discussions: Because a robust system must be able to handle unanticipated events, is there a tradeoff between autonomy and predictability? As autonomy is introduced into a system, are we introducing more software (which usually is accompanied with more bugs), thereby reducing robustness? What are the advantages of introducing autonomy from the beginning (answer: efficiency, capability) versus retrofitting an additional capability (answer: minimal modification to a working system). (Fasciano, 1996).
Experience indicates that autonomous systems find difficulty with user acceptance. We as a community must overcome this roadblock. Proposed steps include (1) involving users in the design process; (2) relying more on demonstrations of the technology versus technical papers to solicit understanding and buy-in; (3) gaining user-pilot trust through long-term use; (4) opening up the "black box" by providing concise explanations of the system's behavior; and (5) allowing for gradual autonomy integration, the level of which can be controlled by the user. (Davis, 1999). No workshop would be complete without a discussion of terminology. As expected, definitions of robustness were plentiful: the ability to produce an executable plan despite run-time variations in state, resources, and activity duration; graceful degradation; comprehensive event or fault detection, identification, and response; and identification of operational bounds or ranges for the system. There is also a level of robustness provided by software engineering approaches, such as understanding all requirements at the beginning of system development versus developing autonomous capabilities and then reengineering them to meet the domain requirements.
In general, all agreed that a helpful step forward would be the definition of a clear set of metrics to determine the level of robustness. Some suggested metrics are to (1) determine how much error-safety margin is accommodated by the system, (2) borrow existing metrics such as the hardware-based approach of identifying the mean time between failure of the system, and (3) develop a generic simulation to benchmark the robustness of autonomous systems. Would metrics such as these provide sufficient evidence to trust a system if your life depended on it? (Fasciano, 1996). From the AI researcher perspective, the increasing realism in computer games makes them an attractive alternative to both robotics in the real world and homegrown simulations.
By working in simulation, researchers interested in human-level AI can concentrate on cognitive capabilities and finesse many of the pesky issues of using real sensor and real motor systems; they must still include some sensor modeling to get realistic behavior, but they don't have to have a team of vision researchers on their staff. They can pursue AI research in worlds that are becoming increasingly realistic simulations of physical and social interactions, without having to create these worlds themselves. (Takahashi, 2000). Computer games are cheap ($ 49. 95), reliable, and sometimes surprisingly accessible, with built-in AI interfaces. Moreover, computer games avoid many of the criticisms often leveled against simulations. They are real products and real environments on their own that millions of humans vigorously interact with and become immersed in. Finally, unlike military simulations, we do not need to hunt out experts on these games; they surround us.
Virtual Reality and applications of artificial intelligence also will play a tremendous role in human lives. VR is a new human-computer interaction in which users are no longer simply external observers of images on a computer screen, but are active participants within a computer-generated 3 -D virtual world. It was pointed out that VR environments differ from traditional displays in that computer graphics and various display and input technologies are integrated to give the user a sense of presence, or immersion, in the virtual environment. The most common approach to the creation of a virtual environment is to fit the user with a head-mounted display. (Bates, 1992).
The sensory aspects of the virtual environment (auditory, visual, and tactile) are delivered to the individual through the head-mounted display. The sensory experiences depend upon the individual's movements within the environment, which are relayed back to the computer from the helmet sensor and other control devices (e. g. , joy stick, the data gloves, or body suit). In short, VR integrates real-time computer graphics, body-tracking devices, visual displays, and other sensory input devices to immerse a participant in a computer-generated virtual environment. (Bates, 1992).
Therapeutic recreation specialists should give special attention to VR applications in health care settings and proactively examine their possible use in a therapeutic recreation intervention program. Words: 3, 348. Bibliography: Atkin, M. S. ; Westbrook, D. L. ; and Cohen, P.
R. 1999. Capture the Flag: Military Simulation Meets Computer Games. In Papers from the AAAI 1999 Spring Symposium on Artificial Intelligence and Computer Games, 1 - 5. Technical Report SS- 99 - 02. Menlo Park, Calif. : AAAI Press. Bates, J. 1992.
Virtual Reality, Art, and Entertainment. Presence: The Journal of Teleoperators and Virtual Environments 1 (1): 133 - 138. Davis, I. 1999. Strategies for Strategy Game AI.
In Papers from the AAAI 1999 Spring Symposium on Artificial Intelligence and Computer Games, 24 - 27. Technical Report SS- 99 - 02. Menlo Park, Calif. : AAAI Press. Fasciano, M.
J. 1996. Real-Time Case-Based Reasoning in a Complex World. Technical Report, TR- 96 - 05, Computer Science Department, University of Chicago. Laird, J.
E. 2000 a. Bridging the Gap between Developers and Researchers. Game Developer 7 (8): 34. Laird, J.
Human-Level AI's Killer Application Interactive Computer Games. AI Magazine, Summer, 2001. Mactaggart, M. No; false starts for IT. Computer Weekly, July 25, 2002.
Mechner, D. All systems go. The Sciences, Jan-Feb, 1998. Provetti, A. & Tran C. AAAI 2001 Spring Symposium Series Reports. AI Magazine.
Fall, 2001. Schaeffer, J. A Gamut of Games. AI Magazine, Fall, 2001. Takahashi, D. 2000. Artificial Intelligence Gurus Win Tech-Game Jobs.
The Wall Street Journal, March 30, 2000, B 14. Time, M. ; Johnson, W. L. ; Jones, R. M. ; Koss, F. ; Laird, J.
E. ; Rosenbloom, P. S. ; and Schwamb, K. 1995. Intelligent Agents for Interactive Simulation Environments. AI Magazine 16 (1): 15 - 39. Whatley, D. 1999. Designing around Pitfalls of Game AI.
Paper presented at the Game Developers' Conference, 15 - 19 March, San Jose, California. Wilson, R. Playing Handset Games. Electronic News, May 28, 2001.
Free research essays on topics related to: menlo park, artificial intelligence, deep blue, brute force, virtual reality
Research essay sample on Artificial Intelligence Deep Blue
