. Jihie Kim and Yolanda Gil. Deriving Expectations to Guide Knowledge Base Creation. In Proc. of the Sixteenth National Conference on Artificial Intelligence,235-241, Menlo Park, CA: AAAI Press, 1999.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....than V 2 specifying how other, factual knowledge should be used. The idea has been that by having a large library of problem solving methods, we can process task independent factual knowledge by applying the right methods to it. There has also been some work to simplify the acquisition of PSMs [28]. Several overview papers [2, 36] contain a detailed classi cation of and literature pointers to many implemented systems. 2.2 Forming expectations from existing knowledge New ways to leverage previously acquired knowledge to guide future acquisition is a prime contribution of our work. We ....

....it for how they t together . If there were some inconsistencies or some, but not enough information was present to run the system, it would ask the expert to x its rules. It was hard coded to nd places were rules could be xed and suggest those to the expert. Expect Method Developer (EMeD) [28] helps an expert contributor create problem solving methods for the expect system. The system works by analyzing interdependencies between methods to create expectations about what other methods need to be de ned and what they will look like. Because methods have stated capabilities which are ....

J. Kim and Y. Gil. Deriving expectations to guide knowledge base creation. In AAAI/IAAI, pages 235-241, 1999.

....actions, the ontology of engineering techniques, the methods for time estimation, and the planner s augmented engineering actions. Since each of these ontologies required different expertise, different knowledge engineers were resposible for each of them. The EXPECT knowledge acquisition tools [Kim and Gil, 1999] were very useful in pointing out inconsistencies between the methods and some of the ontologies, but the planner s actions were not accessible to EXPECT. Consistency across the different parts of the knowledge base had to be maintained largely by hand. Some of the performance problems during ....

Jihie Kim and Yotanda Git. "Deriving Expectations to Guide Knowledge Base Creation." Proceedings of the 6th National Conference on Artificial Intelligence (AAAI-99), pp. 235-241, AAAI/MIT Press, July 18-22 1999.

....originally envisaged by the developer. However, it is difficult for users who are not programmers to add procedural knowledge to systems. In the next section we discuss some of the challenges that users face in more detail. Some KA approaches use expectations of the entered knowledge to aid users [Kim Gil, 1999] . Expectations are beliefs about the knowledge that the user is currently entering that can be used to constrain the possibilities for what can be entered next. For example, expectations can govern the return type of a function that the user is entering, or its general purpose within the system. ....

....problem solving knowledge to be inspected and changed by a user. In their approach, a partially completed KBS can be analyzed to find missing problem solving knowledge that forms the roles to be filled. This is done as part of the interdependency analysis performed by EXPECT [Swartout Gil, 1995; Kim Gil, 1999] which looks at how both problem solving knowledge and factual knowledge is used in the intelligent system. This work extended the role limiting approach to acquire problem solving knowledge and to determine the roles dynamically. However, Gil and Melz s tools were not adequate for end users. ....

[Article contains additional citation context not shown here]

Kim, J., and Gil, Y. 1999. Deriving expectations to guide knowledge-base creation. In Proc. Sixteenth National Conference on Artificial Intelligence, 235-- 241. AAAI Press.

....(There are other kinds of IM including interdependencies between factual knowledge and procedural knowledge. IM can point out missing pieces in solving a problem and predict what pieces are related and how. It has been successfully used in building and checking problem solving knowledge in EXPECT[Kim Gil, 1999; Kim Gil, 2000] To guide users in developing process models, KANAL builds interdependencies among KB objects in the system, and uses them to perform two kinds of checks: static checks and dynamic checks. Static checks are performed by posing questions about various features of the process ....

Kim, J. & Gil, Y. Deriving expectations to guide knowledge base creation. In Proceedings of AAAI-99, pp. 235--241, 1999.

..... We are also investigating the use of KANAL to help teachers formalize process models that will be used as lessons by a tutoring system. 4 Using Interdependency Models In past work, InterdependencyModels have been successfully used in building and checking problem solving knowledge in EXPECT [Kim Gil, 1999; Kim Gil, 2000] They have been used in analyzing how individual components of a knowledge base are related and interact when they are used during problem solving. An example of interdependency between two pieces of procedural knowledge is that one may be used by the other to achieve a subgoal. ....

Kim, J. & Gil, Y. Deriving expectations to guide knowledge base creation. In Proceedings of AAAI-99, pp. 235--241, 1999.

....interfaces need to have many intelligent capabilities in order to support the complex dialogues that they must conduct with the user, integrate the new knowledge with existing knowledge, and make appropriate generalizations. In past research, we developed several acquisition interfaces [10, 2, 18], all using EXPECT as an underlying framework for knowledge representation and reasoning [17] Each interface addressed different issues and helped the user in different ways as they add knowledge to a system, yet none could individually claim to be able to support a user appropriately. This paper ....

....that they are providing knowledge that is useful to the system and whether they have given the system enough knowledge to do something on its own. Our approach is to use Interdependency Models that capture how the individual pieces of knowledge provided work together to reason about the task [10]. These Interdependency Models are derived automatically by the system, and are used to detect inconsistencies and missing knowledge that turn into followup questions to the user. Users are often not sure whether they are on the right track even if they have been making progress, and we have ....

[Article contains additional citation context not shown here]

J. Kim and Y. Gil. Deriving expectations to guide knowledge-base creation. In Proc. Sixteenth National Conference on Artificial Intelligence, pages 235--241. AAAI Press, 1999.

....know that they are providing knowledge that is useful to the system and whether they have given the system enough knowledge to do something on its own. Our approach is to use Interdependency Models that capture how the individual pieces of knowledge provided work together to reason about the task (Kim Gil 1999). These Interdependency Models are derived automatically by the system, and are used to detect inconsistencies and missing knowledge that turn into follow up questions to the user. Users are often not sure whether they are on the right track even if they have been making progress, so we have found ....

....Models to acquire procedural knowledge A theme of our KA research has been how KA tools can exploit Interdependency Models (Swartout Gil 1995) that relate individual componentsof the knowledge base in order to develop expectations of what users need to add next. EMeD (EXPECT Method Developer) (Kim Gil 1999; 2000) a knowledge acquisition tool to acquire problemsolving knowledge, exploits the Interdependency Models to guide users by helping them understand the relationships among the individual elements in the knowledge base. Our hypothesis is that Interdependency Models allow users to enter more ....

Kim, J., and Gil, Y. 1999. Deriving expectations to guide knowledge-base creation. In Proc. Sixteenth National Conference on Artificial Intelligence. AAAI Press.

....new approaches to develop KA tools, in many cases targeted Copyright c #2000, American Association for Artificial Intelligence (www.aaai.org) All rights reserved. to end users though in practice motivated by knowledge engineers. Few user studies have been conducted (Yost 1993; Tallis Gil 1999; Kim Gil 1999), and the participants are typically knowledge engineers. Without studies of the effectiveness of KA approaches and tools for end users, it is hard to assess the actual requirement of end users and our progress towards satisfying them. One of the challenges of this work is to devise a methodology ....

....constraint, the KA tool has the expectation that the user should specify possible fixes, because there is an interdependency between the problemsolving knowledge for finding fixes for violated constraints and the definitions of constraints and their possible fixes. EMeD (EXPECT Method Developer) (Kim Gil 1999), a knowledge acquisition tool to acquire problem solving knowledge, exploits Interdependency Models to guideusers by helping them understand the relationships among the individual elements in the knowledge base. The expectations result from enforcing constraints in the knowledge representation ....

[Article contains additional citation context not shown here]

Kim, J., and Gil, Y. 1999. Deriving expectations to guide knowledge base creation. In Proceedings of the Sixteenth National Conference on Artificial Intelligence, 235--241.

....draw adequate generalizations. We are investigating an alternative and complementary approach to develop acquisition interfaces for problemsolving knowledge that enables users to specify new knowledge directly by using a knowledge editor and associated tools. This paper reports our work on EMeD [Kim and Gil 1999], an acquisition interface that supports users in adding problem solving knowledge to a knowledge based system. The tool helps users understand the relationships among the individual elements in the knowledge base, keep track of missing knowledge that still needs to be added, suggest potential ....

....about interfaces for editing procedural knowledge. The results should be relevant to researchers of end user programming, knowledge based systems environments, and intelligent interfaces at large. EMeD: A METHOD ACQUISITION INTERFACE FOR THE EXPECT ARCHITECTURE EMeD (EXPECT Method Developer) [Kim and Gil 1999] is a knowledge acquisition interface that allows users to specify problem solving knowledge within the EXPECT framework [Gil 1994; Swartout and Gil 1995; Gil and Melz 1997] This section provides a short overview of EMeD, more details can be found in [Kim and Gil 1999] When users add new ....

[Article contains additional citation context not shown here]

Kim, J. and Gil, Y. (1999) Deriving expectations to guide knowledge base creation. In Proceedings of AAAI-99.

....a multi step modification to a KBS and the use of background knowledge about generic tasks. Here we focus on EXPECT s representation of tasks and subtasks. Details of the overall reasoning and knowledge acquisition tools can be found in (Swartout Gil 1995; Gil Melz 1996; Tallis Gil 1999; Kim Gil 1999; Blythe Gil 2000) The problem solving knowledge of an EXPECT KBS consists of a set of methods. Each method has a capability that declares what task can be achieved by the method, a body that describes how the capability is achieved and a return type that characterizes what the method ....

.... fms##### Because it uses structured representations of method capabilities, EXPECT can reason about how different methods relate to each other. This is useful for organizing method libraries (Swartout, Gil, Valente 1999) as well as supporting the acquisition of new problem solving methods (Kim Gil 1999). We mentioned earlier that the representations were developed to support natural language paraphrasing to support explanation generation. This can support knowledge acquisition (KA) from domain experts, since the paraphrase of a capability can be generated automatically from the computer ....

Kim, J., and Gil, Y. 1999. Deriving expectations to guide knowledge-base creation. In Proc. Sixteenth National Conference on Artificial Intelligence.

No context found.

Kim, J. & Gil, Y. #1999a#. Deriving expectations to guide knowledge base creation. In Proceedings of the Sixteenth National ConferenceonArti#cial Intelligence.

No context found.

. Jihie Kim and Yolanda Gil. Deriving Expectations to Guide Knowledge Base Creation. In Proc. of the Sixteenth National Conference on Artificial Intelligence,235-241, Menlo Park, CA: AAAI Press, 1999.

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