Showing papers in "Knowledge Based Systems in 1993"

Wizard of Oz studies — why and how

Abstract: Current approaches to the development of natural language dialogue systems are discussed, and it is claimed that they do not sufficiently consider the unique qualities of man-machine interaction as...

Adaptive systems: from intelligent tutoring to autonomous agents

Abstract: Computer systems which can automatically alter aspects of their functionality or interface to suit the needs of individuals or groups of users have appeared over the years in a variety of guises. Most recently, attention has focused on intelligent interface agents, which are seen as specialised, knowledge-based systems acting on behalf of the user in some aspect of the interaction. Similar requirements for automatic adaptation have been noted in intelligent tutoring systems, natural-language systems and intelligent interfaces. The paper brings together the research which has emanated from a number of backgrounds, and provides a unifying perspective on adaptive systems in general. An architecture for adaptive systems and a methodology for their development are presented. The paper also describes software support for producing adaptive systems, and offers some experimental evidence to justify both the desirability and feasibility of exploiting an adaptive system approach to human-computer interaction

User-centered approach to adaptive interfaces

Abstract: Both adaptive and adaptable user interfaces are intended to fit the needs of individual users and their tasks better. A problem with these interfaces is that users must have and use additional knowledge, either to understand the automatic, system-driven adaptations or to adapt the interface on their own. Beyond these two extreme approaches, an automatically self-adapting system and a user manually adapting the system, the approach introduced in this paper strives for a third option, in which users are able to tell the system how to adapt itself. Accordingly, methods include adaptive adaptation (i.e. modifying the adaptation strategies) and implicit adaptation (i.e. changing the underlying models). The applicability of these findings is illustrated in the context of an adaptive action prompting environment.

Price of flexibility in intelligent interfaces

Abstract: Flexibility and customizability are central to the perceived advantages of the growth in technological powers. However, these and other vectors of technological change, when used clumsily, create new burdens and complexities for beleaguered human practitioners responsible for achieving goals within some field of activity. 'Intelligent interfaces' are sometimes seen as solutions to the growing demands of highly technological and highly automated fields of activity. However, data from a variety of sources indicates that an exclusively technology-driven approach to the development of intelligent interfaces is likely to provide the illusion of assistance while creating a new layer of burdens and complexities. It is necessary to understand the difference between two types of flexibility in cognitive artifacts: (a) flexibilities that increase practitioners' ranges of adaptive responses to the variability in the field of activity, and (b) flexibilities that simply create new burdens on practitioners, especially at high tempos or high criticality periods.

How can knowledge-based systems solve large-scale problems?: model-based decomposition and problem solving

Abstract: A knowledge-based system that can solve large-scale problems is discussed. It is provided with a new method which combines model building with hierarchical problem decomposition. The model-based approach to problem solving is a general method which is ordinarily used by humans. With this method, representations of the problem and the problem-solving process can be separated. Thus, computers provided with this method allows the end user to focus his/her attention on the description of the problem, without paying attention to the description of the method for solving this problem by computer. When a problem becomes large and complex, however, it becomes difficult to solve by this simple model-based method, and it must be combined with another technique to decompose the problem into a set of smaller ones, each of which is simple enough to deal with. This does not assume that the original problem can be decomposed into the same components, because, in general, a problem involves heterogeneous components. A prototype of a system which could implement this idea has been developed, and its effectiveness has partly been proved. Experimentation is going to be carried out on the total system. The paper discusses a method of implementing the idea with the example of representation.

Towards automatic evaluation of multimodal user interfaces

Abstract: The evaluation of the usability and the learnability of a computer system may be performed with predictive models during the design phase. It may be done on the executable code as well as by observing the user in action. In the latter case, data collected in vivo must be processed. The goal is to provide software supports for performing this difficult and time consuming task. The paper presents an early analysis of, and experience relating to, the automatic evaluation of multimodal user interfaces. With this end in view, a generic Wizard of Oz platform has been designed to allow the observation and automatic recording of subjects' behavior while they interact with a multimodal interface. It is then shown how recorded data can be analyzed to detect behavioral patterns, and how deviations of such patterns from a data-flow-oriented task model can be exploited by a software usability critic.

Reasoning by coordination: an integration of case-based and rule-based reasoning systems

Abstract: A case-based reasoning system (CBRS) is appropriate for an experience-rich domain, while a rule-based system performs reasonably well in a knowledge-rich application environment. A CBRS mainly uses past experiences as problem-solving tools, and therefore its capability can be limited when previous experiences are not a good representation of the whole population. On the other hand, a rule-based reasoning system needs a well constructed domain theory as its reasoning basis, and it does not make use of the knowledge embedded in past experiences. In the paper, a multiagent cooperative reasoning system (MCRS), which integrates an inductive-reasoning agent (case-based reasoning system) and a deductive-reasoning agent (rule-based reasoning system), is proposed to solve problems through the cooperation of both agents. The architecture and the inference mechanism of an MCRS are discussed in this paper. A personnel-evaluation support system, based on the proposed MCRS architecture and implemented in prolog, is presented to demonstrate the feasibility of using the MCRS to solve problems in a knowledge-poor and experience-poor domain.

Cognitive task analysis and interface design in a technical troubleshooting domain

Abstract: A model of the interface design process is proposed that uses two interdependent levels of cognitive analysis: (a) study of the criterion task through the analysis of expert/novice differences, and (b) evaluation of the working user interface design through the application of practical interface analysis methodology (the GOMS model). This dual analysis is reviewed in the context of HYDRIVE, an intelligent tutoring system that has been designed to facilitate the development of aircraft hydraulics systems troubleshooting skills. Initial cognitive task analyses identified critical troubleshooting skills and procedures. However, even with an initial cognitive task analysis, the GOMS analysis resulted in significant and beneficial design changes.

Distributed intelligent systems: issues, perspectives and approaches

Abstract: Distributed artificial intelligence is the subarea of AI that uses multiple processors. It has also been referred to as (cooperative) distributed problem solving. Cooperation is necessary because no node has sufficient expertise, information or resources to solve the given problem. DAI studies intelligent coordination. There are several approaches to the achievement of coordination. The paper discusses distributed intelligent systems, and compares them with distributed processing and single agent problem solving systems. The systems are classified on the basis of different strategies for cooperation, and an algebraic style notation is proposed to describe a DIS precisely. This notation can also be used as input to or output from a machine. Certain essential features common to all DISs are also identified.

Design, implementation and evaluation of an adaptive user interface

Abstract: Software systems developed in recent years are becoming increasingly powerful, but in most cases they tend to abandon the user to deal with the complexity of the system alone. There is an immense need for systems with individual, context-sensitive support. The paper discusses the development of such a support system with special focus on the user interface. The extension of an existing and complex program. Excel, to an adaptive and customizable system is described. This extension is called Flexcel, and it has been implemented and evaluated. The primary aim of this extension is to provide a convenient environment for adapting Excel's user interface to particular users and their current tasks. For example, it allows the user to define new menu for actions that are frequently needed and normally require a tedious dialog step. Flexcel also analyzes the user's interaction style and presents adaptation suggestions. A critique component supports the user in his/her problem solving and learning activities related to the adaptability of Flexcel's user interface.

From user interface design to the support of intelligent and adaptive interfaces: an overhaul of user interface software infrastracture

Abstract: Research on adaptive interfaces in the past has lacked support from user interface tools which allow interfaces to be easily created and modified. Also, current user interface tools provide no support for user models which can collect task-oriented information about users. Developing an adaptive interface requires a user model and an adaptation strategy. It also, however, requires a user interface which can be adapted. The latter task is often time-consuming, especially in relation to more sophisticated user interfaces. The paper presents a user interface design environment, UIDE, which has a different software infrastracture. Designers use high-level specifications to create a model of an application and links from the application to various interface components. The model is the heart of all the design and run-time support in UIDE, including automatic dialog sequencing and help generation. UIDE provides automatic support for collecting task-oriented information about users, by the use of its high-level specifications in its application model as a basic construct for a user model. Some examples of adaptive interfaces and adaptive help are presented that use the information that is collectable in UIDE.

Knowledge-based system requirements

Abstract: There is a growing demand both for greater 'ease of use' and 'support in use' as existing software packages are employed for more complex tasks. End users are faced with both the intricacies of the existing software, and the increasing complexity of the application-problem domains. As a result, there are important implications for the design of support systems for the professional user. The paper describes an approach to the identification of user requirements for knowledge-based front ends to existing software systems. The development of KBFEs is a software re-use technique that involves the integration of existing systems and services to provide specific users and user groups with a tailored solution. As a result, a KBFE enables a greater number and variety of users to have access to existing software and its functionality. This has implications for the requirements-elicitation methods employed. Investigations requirements that assume the use of existing software are described, and a framework for the structuring of the user requirements obtained is presented.

Knowledge-based system for computer-aided drug design

Abstract: The paper presents DD-CHAUS, a drug design aiding system based on a logic-programming approach. One of its characteristics is that it addresses the combinatorial-explosion problem. This has been realized in two ways: representing design strategies as meta rules (DP_META), and simulating the design examples of the past according to the analogy between chemical graphs (DP_ANALOGY). A special format to describe knowledge of drug design, EMIL DATA SHEET, is used to acquire expert knowledge. Compared with other drug design aiding systems, DD-CHAUS is more flexible in terms of the description, management and utilization of knowledge and strategies.

Efficient hypothetical reasoning system for predicate-logic knowledge base

Abstract: A hypothetical reasoning system is an important framework in the development of advanced knowledge-based systems. It can be effectively applied to many practical problems including model-based diagnosis, and designs. However, the inference speed of its prolog -based implementation is slow, and this is particularly because of inefficient backtracking. In order to overcome this problem, a fast hypothetical reasoning mechanism for propositional-logic knowlegde has been developed by combining the advantages of forward and backward reasoning styles. This fast mechanism, however, cannot be applied to hypothetical reasoning with predicate-logic knowledge where variables are included as arguments. The paper presents a fast hypothetical reasoning mechanism for predicate-logic knowledge as an extension of the above idea. A reasoning method developed in the deductive database area is effectively utilized to realize this fast mechanism, which can even manipulate recursive rules.

From scientific computation to decision support

Abstract: Research institutes and universities have developed many programs which have been used up until now mainly by specialists. The input preparation, the interpretation of the output, and the use of the program are often too complicated to be suitable for a wider group of users. A part of the knowledge necessary for the use of a scientific program can be included in a user interface. The development of such user interfaces is a mixture of the development of an expert system and a normal user interface. The impact of the new user interface on the changing user task has to be taken into account in the development process.

Feature extraction and validation within a flexible manufacturing protocol

Abstract: An intelligent feature extraction methodology has been developed for use with a flexible manufacturing protocol. A wireframe input scheme based on the standard Initial Graphics Exchange Specifications format is used. The methodology has nine tasks. The first task involves the processing of a wireframe CAD file in the IGES format. The last task generates a feature file represented in another standard format. The intermediate tasks are concerned with locating the planes, the contours, the cylindrical surfaces and the boundary faces present within the part drawing. This methodology has been implemented on a Sun microcomputer within the Franz Lizp environment. The feature extractor runs in both automated and interactive modes. It accepts CAD files generated by any commercial CAD tool with an IGES interface. It has been tested with two such tools.

Using the FOCUS architecture for developing knowledge-based front ends: a KBFE for forecasting

Abstract: Statistical packages, optimisation packages and numerical-algorithms libraries are widely used in industrial and scientific development environments. They represent an enormous body of very complex and valuable knowledge that is becoming increasingly difficult to access. End users of these systems have to cope simultaneously with the intricacies of the software and with the increasing complexity of the application-domain problems. For these systems, knowledge-based front ends can provide co-operative assistance to end users, enabling them to use the systems successfully, while preserving the knowhow contained in the libraries and packages. They are also valuable in extending the software's working life. FOCUS is an ESPRIT-2 project whose goal is to develop generic tools and techniques for constructing and maintaining KBFEs for open user systems (e.g. libraries, reusable software components) and closed user systems (e.g. free-standing software, packages) for industrial and scientific applications. The participating partners are drawn from both industrial and academic institutions, providing a wide crosssection of software researchers, producers and users, and, although there are substantial research objectives, the project has taken a pragmatic approach, with the commercialisation of products developed playing a prominent role. The paper describes part of the work undertaken during the twenty-fourth month of the ESPRIT II project, which began in December 1988 and ran for four years. The aims of the FOCUS project are stated, and a core element of the emerging FOCUS design strategy is described, namely a separable architecture for knowledge-based front ends, concentrating on the front-end harness, the problem solver and the back-end manager components. Some KBFEs developed using this architecture, and the tools created within the project that operate in the domain of forecasting, are also described.

Reachability trees for Petri nets: a heuristic approach

Abstract: The paper explores the use of heuristics to search reachability trees (graphs) for particular markings. Evaluation functions are defined on markings for two well known heuristics: means-ends analysis and the A ∗ algorithm. A comparison of these approaches with conventional reachability tree analysis (breadth-first search) is given.

Design of an intelligent interface to standard PC applications which maximises the ability of the disabled user

Abstract: The paper reports on work on the design of speech and switch-driven interfaces. It presents a design for a generic user-interface management system (UIMS) which allows the user access to any PC application via both speech and switched input. This generic UIMS allows an instantiation of an interface which is tailored to the requirements of a particular individual, and which can, if necessary, be changed over time as his/her condition changes. Interaction is aided by a knowledge of the user's requirements at the task level, drawing on interaction history and the knowledge of the application gained from a generic task model. A user-oriented task model is derived during the interaction, thus providing enhanced performance over time.

Can complex conventional programs be revitalized

Abstract: The paper describes a knowledge-based front end for a mathematical simulation program CHEMSIM that has become outdated because of the evolution of interface techniques. Reimplementation is out of the question because of the program's complexity. The front end helps the user with its data entry. The task model and user characteristics specified the functional description. A hierarchical task tree models the domain knowledge and user interface. A free walk through the parameters is guided by graphical representation and known technical graphics. The front end takes care of consistency and the inter-connections of parameters, offers context dependent help, and gives feedback on the user's progress. Reusability was one important objective during the development phases.

Towards competent information acquisition interactions between an expert system and its user

Abstract: Expert systems are consultative, highly interactive systems, and hence the quality of interaction between system and user is important for the acceptability of the system. Acquisition of data about the problem in hand is a central feature of the interaction between system and user, and it makes a major contribution to the user's perception of the system. It is therefore crucial that the acquisition of this data (both as individual data items and as a sequence of data inputs) is perceived by the user as competent. This paper identifies central, domain-independent, design goals for the information-acquisition interactions between an expert system and its user, including mixed-initiative interaction, flexibility in user input, and system competence in querying the user. The paper discusses the realisation of these goals in a diagnostic expert system, Skeletal Dysplasias Diagnostician, through an explicit data model which allows for the representation of data with temporal and spatial (locality) aspects and the decide-status function which operates on the data model.

Knowledge-based approach to structural integrity verification in requirements analysis and logical system specification

Abstract: Constructing an accurate representation of an information system that meets user requirements and translating these requirements into appropriate logical system specifications are two time-consuming tasks in the systems development life cycle. Though many structured methodologies are effective in enhancing communication between analysts and users, they provide limited support in representing an information system accurately. Furthermore, translating a user requirement, expressed as a function of this representation, into its logical equivalent is still a tedious task that is often performed manually by an analyst. This paper describes a knowledge-based approach to the representation of an information system as a Petri net. The characteristics of the network are then used to verify the system's structural integrity during the requirements analysis and system specification phases. The knowledge-based approach described is implemented in the logic programming language PROLOG.

Multiple context planning within a truth-maintenance system

Abstract: The paper demonstrates the feasibility of planning using the properties of operations available in a factory, without the explicit statement of the relationships between these operations being required. This approach allows the generation of plans to be broken down into sub-problems which can be performed, if necessary, by specialist planners, a framework for integrating specialised planning systems thus being produced.

Plans in natural-language dialogues

Abstract: This paper suggests reflecting the ‘state of mind’ of participants in conversation by viewing plans as a kind of mental structure, a structure that takes into account and affects such mental states as beliefs, commitments and intentions. A formal method is described for computing the construction and the reconstruction of mental plans. Plan reconstruction is then used in the context of the man-machine interface to (a) model topic structure, (b) generate appropriate responses when agents use speech acts, and (c) respond to queries that result from an invalid plan.

FEATS: knowledge representation and processing issues

Abstract: FEATS is an intelligent training system that supervises a structural/mechanical engineer using a finite element package such as COSMIC/Nastran. Within this goal, FEATS involves the investigation of various aspects of artificial intelligence including natural language understanding and processing, goal extraction and planning, user modelling, and knowledge representation and manipulation. The paper describes an approach to understanding the engineer's conversation with FEATS, the consequent extraction of knowledge, and its representation. The knowledge representation and processing is implemented in a FEATS prototype. This prototype was written in Quintus prolog and Quintus ProWindows, and it runs on a Sun SparcStation 1.