Chandrasekaran B., and Johnson T.R., "Generic Tasks and Task Structures: History, Critique and New Directions," in Second Generation Expert Systems, David J.M., Krivine J.P., and Simmons R. (Eds.), Springer-Verlag, London, pp. 233-272. (1993).  Home/Search   Document Details and Download   Summary   Related Articles   Check

.... (as a result of partial agreement) across two ontologies depends on the level of abstraction, i.e. a lack of shareability appears in the deeper knowledge (as discussed above) To improve knowledge reusability, several researchers from the problem solving area (e.g. Chandrasekaran and Johnson [CJ93], Clancey [C92] or Swartout and Moore [SM93] have proposed the idea of structuring the knowledge into different levels of abstractions, while Steels in [S93] proposed a componential framework that decomposes a knowledge level into reusable components. In addition to the level of abstraction, ....

Chandrasekaran B., and Johnson T.R., "Generic Tasks and Task Structures: History, Critique and New Directions," in Second Generation Expert Systems, David J.M., Krivine J.P., and Simmons R. (Eds.), Springer-Verlag, London, pp. 233-272. (1993).

....on separating strategic knowledge, S, from K. It yielded systems like NEOMYCIN, which typically modeled an application task as a hierarchy of procedural subtasks, where high level subtasks capture S and use it to sequence lower level subtasks, all the way to primitive subtasks that directly use K [5]. Capturing S in this fashion helped to make certain control decisions explicit, and thus M. Benaroch Goal Directed Reasoning with ACE SSM 2 2 enabled KBSs to explain some aspects of their behavior. More importantly, many subtasks capturing S became viewed as task independent reusable ....

....that, although the behavior of many known KBSs is characterized as being goaldriven, these systems are not really aware of their goals. This observation applies even to systems like ABLE and NEOMYCIN, in which somewhat explicit SSMs emerge from interactions of the subtasks they use to capture S [5]. This paper respectively focuses on two issues. One is how to make goal knowledge, G, explicit as well. The paper specifically explains how G is represented and used with ACE SSM an architecture for constructing explicit SSMs. The second issue is how availability of G and the explicit SSMs it ....

[Article contains additional citation context not shown here]

Chandrasekaran, B., and Jonhson T.R., "Generic Tasks and Task Structures: History, Critique and New Directions," in Second Generation Expert Systems, David J.M., Krivine J.P., and Simmons R. (Eds.), Springer-Verlag, London, pp. 233-272, 1993.

....should be identified. This can be done from scratch by identifying the agents in an application and their collaborations. Alternatively, a role model catalog can be used; this provides commonly occurring agent role models. A role model catalog resembles a set of patterns or a pattern language [8, 12, 13], and role model catalogs are being used at BT for downstreaming agent technology. The role model catalog developed at BT includes approximately 60 role models, including those from agent enhanced workflow, flexible manufacturing, electronic E. A. Kendall, Role Models for Analysis, Design, and ....

....Manager Bureaucracy Subordinate Bureaucracy AEW Agent Client Bureaucracy Figure 6: Role Composition during Design 4.0 OBJECT ORIENTED IMPLEMENTATIONS OF ROLE MODELS Whenever possible, design patterns should be utilized to design and implement software systems. Several patterns [3, 7, 8] exist for addressing applications where objects play roles; the Role Object pattern will be utilized here. A design based on the Role Object pattern is depicted in Figure 7; it provides an object oriented design (UML notation) for an agent that participates in supply chains and bureaucracies and ....

Chandrasekaran, B., Johnson, T, "Generic Tasks and Task Structures: History, Critique and New Directions", in Krivine, D., & Simmons, R., (Eds.), Second Generation Expert Systems, Springer Verlag, 1993.

....the reuse of (parts of) existing engines. In addition, our platform provides for a corresponding task oriented representation framework, to guide domain experts in expressing their knowledge at the expertise level. Our approach follows well known trends about second generation expert systems [27, 7], and is in the spirit of knowledge engineering environments such as SBF [17] Prot#g# II [13] or Par KAP [24] Lama emphasizes the software level composition of KBS engines (hence PSMs) for a given task from reusable reasoning blocks. These are of smaller grainsize than classical PSMs, such as ....

....to a personal model, be it explicit or not, it is to note that no particular mechanism is included in the platform to help decide how to combine the instructions provided by Lama to implement the appropriate PSM. Many of the ideas discussed here have been already discussed in past literature [27, 17, 7, 13, 3, 14], given as guidelines to develop KBS. Our experience is interesting as it shows how these can be followed for the operationalization of KBSs for a real world task such as PS, from a software engineering point of view. 6 Conclusions and future work In this paper we present a case study of KBS ....

B. Chandrasekaran and T.R. Johnson, `Generic Tasks and Task Structures: History, Critique and New Directions ', in Second Generation Expert Systems, eds., J. David, J. Krivine, and R.Simons, 232272, SpringerVerlag, (1993).

....applications in general) but by allowing one to profit from generic conceptual designs in other ways than currently offered. 3 To appear in IEEE Expert, 1995. Probably the most well known and accepted reusable components for conceptual knowledge system design include those characterized in [1, 2, 3, 4]. Important differences exist but with respect to reusability, these generic conceptual designs share this major idea: The final conceptual knowledge system design (or parts of this design) created in a previous experience are abstracted from the application specific situation. The resulting ....

B. Chandrasekaran and T. Johnson, "Generic tasks and task structures: History, critique and new directions," in Second Generation Expert Systems (J. David, J. Krivine, and R. Simmons, eds.), pp. 232--272, Springer-Verlag, Berlin, 1993.

....about in a manner suitable for a particular class of tasks. Moreover, in knowledge level research it is recognized that interactions exist between task, domain, and inference knowledge. Indeed, domain knowledge cannot be adequately represented independent of the class of tasks it is designed for [3]; Task knowledge is related to inferences through the competence of an inference structure to derive the required outputs from the task s inputs; Domain knowledge is related to inference knowledge through its ability to fill the inference s knowledge roles. Encapsulating problem solving behaviors ....

Chandrasekaran, B. and Johnson, T., "Generic tasks and task structures: History, critique and new directions ". In David, J., Krivine, J., and Simmons, R., editors, Second Generation Expert Systems, pages 232--272. Springer-Verlag, Berlin, 1993.

No context found.

Chandrasekaran B., and Jonhson T.R., 1993, "Generic Tasks and Task Structures: History, Critique and New Directions," in Second Generation Expert Systems, David J.M., Krivine J.P., and Simmons R. (Eds.), Springer-Verlag, London, pp. 233-272.

No context found.

Chandrasekaran, B., Johnson, T, "Generic Tasks and Task Structures: History, Critique and New Directions", in Krivine, D., & Simmons, R., (Eds.), Second Generation Expert Systems, Springer Verlag, 1993.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC