Traditional softwareengineering is crucial to the successful development and deployment of a large expert system. In this respect, large expert systems are shown to have many similarities with other large software sy...
详细信息
Traditional softwareengineering is crucial to the successful development of a large expert system. In this respect, large expert systems are shown to have many similarities with other large software systems. The adva...
详细信息
Traditional softwareengineering is crucial to the successful development of a large expert system. In this respect, large expert systems are shown to have many similarities with other large software systems. The advanced computing concepts, rapid prototyping approach, and general research and development aspects of expert systems have sometimes led to the neglect of fundamental principles of softwareengineering. These principles are creating and maintaining interfaces between program modules, the use of specification languages, automatic code generation, long-term maintenance, and the use of commercial products. A large-scale expert system currently in advanced development is used as an example.< >
The relationship between nonmonotonic formalisms and the problem of temporal projection has long been recognized, though not well understood. The straightforward application of various nonmonotonic formalisms (circums...
ISBN:
(纸本)9780897913720
The relationship between nonmonotonic formalisms and the problem of temporal projection has long been recognized, though not well understood. The straightforward application of various nonmonotonic formalisms (circumscription, default logic, etc.) has been blocked by the multiple-extension problem [Hanks and McDermott 87], where a seemingly intuitive set of axioms gives rise to extensions (or minimal models, depending on the formalism) other than the one desired. However, recent work in the semantics of logic programs has gone a long way toward clarifying this situation. In particular, a class of logic programs (called the stratified logic programs) has been identified in which nonmonotonic reasoning can be carried out in a well-defined fashion [Przymusinski 88].This paper presents a system, based on the Event Calculus [Kowalski and Sergot 86], for temporal reasoning in a deductive database context. The axioms used to perform temporal projection are constructed to satisfy the stratification property, so that nonmonotonic inference can be performed safely. The system employs a caching mechanism which stores conclusions drawn from the axioms along with justifications. The cache is managed using truth maintenance techniques which faithfully reflect the nonmonotonic semantics of the system.
作者：
CALVERT, TERODRIGUEZ, FASLEBZAK, JSThomas E. Calvert
P.E.: is a senior project engineer with the Propulsion and Auxiliary Systems Department David Taylor Research Center Annapolis Md. His interests include application of computers to all aspects of engineering with particular emphasis on utilization of small computers. Mr. Calvert is a licensed professional engineer in Maryland. He received a BSEE from Drexel University in 1969 and since that time has completed a number of graduate courses related to machinery acoustics. Francisco A. Rodriguez:is an engineer with the Propulsion and Auxiliary Systems Department
David Taylor Research Center. He was formerly with the Computer-Aided Design/Interactive Graphics Group of the Division of Engineering and Weapons U.S. Naval Academy. His interests include interfacing the computer aided design to the computer aided manufacturing along with related software and hardware development. Mr. Rodriguez received a BSEEfrom the University of Virginia in 1968. James S. Slebzak:is a mechanical engineering technician with the Propulsion and Auxiliary Systems Department
David Taylor Research Center. He received his machinists papers in 1971 after serving his apprenticeship at David Taylor Research Center. He continued his education and became the senior numerical control programmer at the Annapolis Laboratory. He completed his mechanical technology degree from Anne Arundel Community College in 1986. His interests are in the application of numerically controlled manufacturing techniques to prototype machinery components.
The machine shop at the Annapolis Laboratory of the David Taylor Research Center (DTRC) provides model making and prototype support to a large variety of naval ship related engineering projects. In order to meet these...
详细信息
The machine shop at the Annapolis Laboratory of the David Taylor Research Center (DTRC) provides model making and prototype support to a large variety of naval ship related engineering projects. In order to meet these challenging requirements, computer aided design/computer aided manufacturing (CAD/CAM) techniques are being used to produce “one of a kind” prototypes or very low volume production parts. The use of computer aides in these cases is to facilitate the accurate manufacture of a difficult part, rather than to improve manufacturing efficiencies. In addition, the approach provides the flexibility required to support research and development projects. Several examples of prototype shipboard components manufactured using CAD/CAM techniques are presented in this paper. The hardware and software that facilitated these projects are discussed. The examples described have met the requirements to produce a wide variety of prototype shipboard machinery components quickly and accurately.
In this paper, we present a logic-based theory for the performance of two types of commonsense reasoning: inheritance and temporal reasoning. The inheritance part of the theory is modeled on frame-based systems, while...
详细信息
Summary form only given, as follows. A neural network speech recognition system was constructed based on a feedback network with backpropagation in an effort to explore various issues relating to representation, scali...
详细信息
Summary form only given, as follows. A neural network speech recognition system was constructed based on a feedback network with backpropagation in an effort to explore various issues relating to representation, scaling, generalization, and time invariance. The network was trained on stop consonant and vowel data obtained from continuous speech. The results indicated a surprising tendency for the network to construct local rather than distributed representations. The stability of learning increased with network size, and generalization was not impaired in large scale networks. Increased complexity of the recognition problem did reduce recognition performance and learning stability, and increase learning time. The network showed capabilities of learning to recognize speech sounds that were not synchronized to a particular time.< >
Achieving maximum performance through migration of functions from software to microcode requires rethinking the linkage editing process. An object-oriented model of naming and binding clarifies the alternative abstrac...
详细信息
Achieving maximum performance through migration of functions from software to microcode requires rethinking the linkage editing process. An object-oriented model of naming and binding clarifies the alternative abstractions available in naming and linking across the macro-micro machine boundary. Alternative abstractions for sharing micro-objects and for dynamic use of micro-objects are presented and their implementations discussed. The conclusions are based on actual implementations.
This paper examines the historical oscillations of artificial intelligence (AI). It traces the very beginnings from 1956 at the Dartmouth Summer Research Project on Artificial Intelligence, through the depressions of ...
详细信息
This paper examines the historical oscillations of artificial intelligence (AI). It traces the very beginnings from 1956 at the Dartmouth Summer Research Project on Artificial Intelligence, through the depressions of the 1960s and 1970s, to the exponentially increasing number of successful applications in the mid-1980s. It shows how to select problems which can be solved with the aid of expert systems. The paper also identifies areas such as common sense reasoning, automated machine learning, and complex design synthesis which are now beyond the state of the art and will be for years to come. Artificial intelligence programs are able to diagnose faults and classify solutions in narrowly defined specialties even when the data is “fuzzy,” but they have not exhibited autonomous “thinking.” Just as conventional computer programming has alleviated the burden of calculating, AI expert systems will streamline the processing of logical data. Both of these computer techniques are cost effective when they are applied to well defined tasks, since computers are faster than people and error free for routine tasks.
DynamicDesign is a computer-Aided softwareengineering environment for the C language with a layered system architecture for modularity and versatility. DynamicDesign is composed of facilities to edit hypertext object...
ISBN:
(纸本)9780897913409
DynamicDesign is a computer-Aided softwareengineering environment for the C language with a layered system architecture for modularity and versatility. DynamicDesign is composed of facilities to edit hypertext objects, maneuver thorough hypertext graphs, build a hypertext graph from a set of existing C source files, and browse source code, documents and system requirements. This paper discusses the DynamicDesign facilities that deal with the source code, sourceBrowser, and source tree builder *** is a utility used to convert C source code into a hypertext source graph, based on the program's call tree. A data dictionary is constructed for the program that contains its local and global *** source browser allows the user to traverse, view, and edit a source code tree. Additional facilities for understanding and maintaining the source code and its auxiliary documentation are provided by the browser.
The Hypertext Abstract Machine (HAM) is a general-purpose, transaction-based, server for a hypertext storage system. The server is designed to handle multiple users in a networked environment. The storage system consi...
ISBN:
(纸本)9780897913409
The Hypertext Abstract Machine (HAM) is a general-purpose, transaction-based, server for a hypertext storage system. The server is designed to handle multiple users in a networked environment. The storage system consists of a collection of contexts, nodes, links, and attributes that make up a hypertext graph. This paper demonstrates the HAM's versatility by showing how Guide1 buttons. Intermedia webs, and NoteCards FileBoxes can be implemented using the HAM's storage model.
暂无评论