Use Of Expert Knowledge Research Articles

Solving black box computation problems using expert knowledge theory and methods

The challenge problems for the Epistemic Uncertainty Workshop at Sandia National Laboratories provide common ground for comparing different mathematical theories of uncertainty, referred to as General Information Theories (GITs). These problems also present the opportunity to discuss the use of expert knowledge as an important constituent of uncertainty quantification. More specifically, how do the principles and methods of eliciting and analyzing expert knowledge apply to these problems and similar ones encountered in complex technical problem solving and decision making? We will address this question, demonstrating how the elicitation issues and the knowledge that experts provide can be used to assess the uncertainty in outputs that emerge from a black box model or computational code represented by the challenge problems. In our experience, the rich collection of GITs provides an opportunity to capture the experts' knowledge and associated uncertainties consistent with their thinking, problem solving, and problem representation. The elicitation process is rightly treated as part of an overall analytical approach, and the information elicited is not simply a source of data. In this paper, we detail how the elicitation process itself impacts the analyst's ability to represent, aggregate, and propagate uncertainty, as well as how to interpret uncertainties in outputs. While this approach does not advocate a specific GIT, answers under uncertainty do result from the elicitation.

Whom to Trust? Public Concerns, Late Modern Risks, and Expert Trustworthiness

This article discusses the conditions under which the use of expert knowledge may provide an adequate response to public concerns about high-tech, late modern risks. Scientific risk estimation has more than once led to expert controversies. When these controversies occur, the public at large – as a media audience – faces a paradoxical situation: on the one hand it must rely on the expertise of scientists as represented in the mass media, but on the other it is confused by competing expert claims in the absence of any clear-cut standard to judge these claims. The question then arises, what expertise can the public trust? I argue that expert controversies cannot be settled by appealing to neutral, impartial expertise, because each use of expert knowledge in applied contexts is inextricably bound up with normative and evaluative assumptions. This value-laden nature of expert contributions, however, does not necessarily force us to adopt a relativist conception of expert knowledge. Nor does it imply active involvement of ordinary citizens in scientific risk estimation – as some authors seem to suggest. The value-laden, or partisan, nature of expert statements rather requires an unbiased process of expert dispute in which experts and counter-experts can participate. Moreover, instead of being a reason for discrediting expert contributions, experts' commitment may enhance public trustworthiness because it enlarges the scope of perspectives taken into account, to include public concerns. Experts who share the same worries as (some of) the public could be expected to voice these worries at the level of expert dispute. Thus, a broadly shaped expert dispute, that is accessible to both proponents and opponents, is a prerequisite for public trust.

Fuzzy time series approach for disruption prediction in tokamak reactors

Disruption in a Tokamak reactor is a sudden loss of confinement that can cause damage to the machine walls and support structures. In this paper, we propose the use of a fuzzy time series (FTS) approach for detection of disruption in Tokamaks. In particular, two-factors FTS models have been exploited for the prediction of the time to disruption in the ASDEX-Upgrade machine. The concept of fuzzy logic is used taking into account that previous techniques make use of expert knowledge for deciding about the onset of a disruption.

Mapping Expert Knowledge: Redesigning the Monarch Butterfly Biosphere Reserve

Abstract: Expert consultation has been used to fill the information gaps that hamper conservation planning and nature reserve design. The use of expert knowledge in conservation planning is difficult, however, because it is subjective, biased, and value‐laden. Decision theory provides a systematic and comprehensive means for addressing experts' subjective—and sometimes contradictory—judgments in the design of nature reserves. Thus, the experts can separate the objective criteria from the subjective components of decision making that place value on those criteria. When linked to a geographic information system ( GIS ), these techniques foster consensus among experts by allowing the exploration of alternative designs in an iterative way. We used such a decision‐analysis approach to redesign the Monarch Butterfly Biosphere Reserve ( MBBR ) in central Mexico. We examined three reserve scenarios to identify the optimal overwintering habitats considering ( 1 ) an area equivalent to the previously defined boundaries of the core zone of the MBBR ( 4500 ha ); ( 2 ) an area equivalent to the previously defined boundaries of the whole MBBR ( 16,000 ha ); and ( 3 ) the maximum possible extent for a new core zone. This last scenario produced an area of 21,727 ha. These results were transferred to the GIS to create the respective nominal maps that were presented to the environmental authorities, who selected the third scenario for the core zone of the new MBBR. Our results allowed us to locate the prime overwintering habitats precisely and to delimit a core area for the reserve that would minimize the inclusion of forest stands valuable to local loggers.

Re-tracing the Path to Third Order Knowledge: How did we get here and can we move on?

The paper argues that policy-oriented educational research, specifically large-scale quantitative research, has produced a new kind of knowledge, which is herein termed third order knowledge. This expert knowledge distinguishes itself from tacit knowledge, everyday knowledge and social science knowledge through its purposes and sources. Third order knowledge is defined as knowledge of and for experts that has as its purpose the legislation and regulation of everyday lives. It is proposed that the power and influence vested in the production and use of expert knowledge motivates a critical analysis. This analysis is contextualized in an historical study of representations of knowledge, using a philosophical framework and social theory.

Profiles of effective tutors in problem-based learning: scaffolding student learning.

Research on tutoring in problem-based learning has not focused so far on the variation in tutoring and how this variation can be interpreted by conceptions about effective tutoring. This study focuses on the profiles of tutors generated by means of an instrument, the so-called Tutor Intervention Profile (TIP), and tries to determine which profiles are more or less effective. The TIP contains four dimensions of tutor behaviour: (1) elaboration; (2) directing the learning process; (3) integration of knowledge; and (4) stimulating interaction and individual accountability. The medical school of the University of Maastricht, The Netherlands. Sixty-seven tutors who run 67 tutorial groups across three units (courses) in the academic year 1996-97. It appeared that high, average and low performing tutors differ in their performance on each of the four dimensions of the TIP. Several different profiles of tutor performance could be distinguished, which were more or less effective. One group of tutors demonstrated a tutor intervention profile that was characterized as relying more on the use of expert knowledge, whereas another group of tutors was characterized as relying more on their abilities to stimulate the learning process in the tutorial group. The tutor intervention profile that was perceived by students as most effective showed high scores on each of the four dimensions, as expected. Notably, a tutor stressing the learning process in the tutorial group was perceived as more effective than a tutor stressing content (expert tutor). This is especially true for a relatively poor scoring tutor. The results of this study are consistent with research on human tutoring and research on tutoring in problem-based learning.

Fuzzy Neural Network for State Estimation and Optimal Control in L-Isoleucine Fermentation

A state pre-estimation model of L-isoleucine fermentation is established by using neural network. The chemotaxis BP algorithm that can effectively overcome the local minimum is adopted for training the net. A neural network that equals a fuzzy system is applied for getting air flow to control dissolved oxygen.the membership functions are determined by the net learning. A new control method based on neural networks and fuzzy logic is proposed. This scheme can fully make use of expert knowledge and is robust for uncertainty in model and accurate with noise immunity.

Method for segmenting chest CT image data using an anatomical model: preliminary results.

We present an automated, knowledge-based method for segmenting chest computed tomography (CT) datasets. Anatomical knowledge including expected volume, shape, relative position, and X-ray attenuation of organs provides feature constraints that guide the segmentation process. Knowledge is represented at a high level using an explicit anatomical model. The model is stored in a frame-based semantic network and anatomical variability is incorporated using fuzzy sets. A blackboard architecture permits the data representation and processing algorithms in the model domain to be independent of those in the image domain. Knowledge-constrained segmentation routines extract contiguous three-dimensional (3-D) sets of voxels, and their feature-space representations are posted on the blackboard. An inference engine uses fuzzy logic to match image to model objects based on the feature constraints. Strict separation of model and image domains allows for systematic extension of the knowledge base. In preliminary experiments, the method has been applied to a small number of thoracic CT datasets. Based on subjective visual assessment by experienced thoracic radiologists, basic anatomic structures such as the lungs, central tracheobronchial tree, chest wall, and mediastinum were successfully segmented. To demonstrate the extensibility of the system, knowledge was added to represent the more complex anatomy of lung lesions in contact with vessels or the chest wall. Visual inspection of these segmented lesions was also favorable. These preliminary results suggest that use of expert knowledge provides an increased level of automation compared with low-level segmentation techniques. Moreover, the knowledge-based approach may better discriminate between structures of similar attenuation and anatomic contiguity. Further validation is required.

An Interactive Dynamic Inventory-Production Control System

The development and evaluation of an interactive model for controlling inventories and scheduling production in a high technology batch production environment are discussed. The model allows the use of expert knowledge and company policy for matters such as customer service and safety stocks, in an interactive way. Production of a large number of products has to be scheduled for a number of time periods and a modified Wagner-Whitin algorithm is used for developing good schedules. A versatile data and information management system is a part of the developed software which has been implemented using a standard microcomputer. An evaluation suggests that the system is a practical one for use by busy managers. Further, it results in substantial cost savings and increased customer service.

Potential for expert systems in the assessment and repair of fire damaged buildings in India

Assessment of structural damage due to fire in buildings can be a complex process plagued by uncertainty. This complexity, argue the authors, arises from subjective opinion and imprecise numerical data. Expert system technology offers significant potential for a systematic assessment and repair of fire damaged buildings through the use of expert knowledge.

The utilization of artificial intelligence to achieve availability improvement in automated manufacture

This paper is intended to outline both the potential and use of artificial intelligence (AI) in the context of failure diagnostics for automated manufacturing systems. The specific objective is to illustrate how improvements in the availability of production systems can be achieved via the use of AI based system diagnostics. Accordingly, the paper describes several on-going application studies in this area, which involve the use of expert knowledge based systems (EKBS) and artificial neural nets (ANN), techniques which are shown to possess a potential for fast and efficient system diagnostics and, thereby, improve the maintenance of manufacturing equipment so as to achieve high availability.

Rule-based adaptive configuration selection for analog design systems

A configuration selection tool for rule-based analog IC design is presented. A fast determination of proper analog circuit configuration plays an eminent role in the initial design phase. This appears to be quite demanding due to the large number of parameters and side issues involved. In order to ease concept engineering in analog design an adaptive algorithm has been developed for configuration selection. It can be used in design systems that rely on use of expert knowledge stored as a set of rules in a knowledge base, e.g., expert systems, design assistants, etc. The developed learning algorithm also helps eliminate inconsistent, incorrect, and redundant rules in rule bases used in analog design systems. The rule base is continuously adapted, depending on the correctness of the decisions the rules have invoked.

What Is Necessary for Proof? Is 95% Sure Unrealistic?

To the Editor. —The discussion by Dr Eddy 1 of what is needed for evidence to be "sufficient" when evaluating medical treatments deserves further exploration. Even if we can agree on appropriate end points, study designs, and use of expert knowledge, the question of how certain we must be that treatment A is beneficial compared with treatment B is still

Template-Based Fuzzy Systems Modeling

A methodology is suggested here for the development of fuzzy systems models that is a combination of the AI-expert systems approach, with its heavy dependence on the use of expert knowledge, and the neural network-type systems building, with its emphasis on learning from data observations. We use expert-provided information in the form of template linguistic values to induce potential elemental rules for the knowledge base of the system model. We then introduce input-output observations into a simple learning mechanism to obtain weights characterizing the effect of each of the potential elemental rules on the overall systems model. The development of the learning mechanism is based on a representation of systems models combining fuzzy logic and Dempster-Shafer theory, which we previously introduced.

A comparison of qualitative and quantitative physical land evaluations, using an assessment of the potential for sugar‐beet growth in the European Community

Abstract. Qualitative and quantitative physical land evaluations differ in their technical procedures, i.e. use of expert knowledge versus process‐orientated simulation models. This paper compares the results of both procedures using the growth potential for sugar‐beet in the European Community. Qualitative procedures give suitability expressions, such as land being well suited or moderately suited for a given land use. Less than 30% of EC land was found to be well suited or moderately suited under water‐limited conditions. No quantitative expressions for the crop yield potential are, however, produced. The quantitative procedure describes suitability in terms of average crop yield and its temporal variability. Water‐limited and potential dry matter yields of sugar‐beet were estimated to range from 3.5 to 20 t/ha, and from 6 to 25 t/ha, respectively.Comparison of results of the qualitative and quantitative procedures for regions showed that the suitability classes obtained by the former can be characterized by different yield distributions derived from the latter. These yield assessments showed that results of qualitative land evaluation procedures aimed at assessing yield potential are meaningful only when they are linked to agroclimatic zones. The comparison also showed that some moderate restrictions, which often can be counteracted by adequate farm management, are not incorporated into the quantitative procedure.

The Bhopal Tragedy: Language, Logic, and Politics in the Production of a Hazard.

This book offers a critical theory of hazards, which the Bhopal tragedy would serve to highlight. It provides a general model of how hazards existed in Bhopal come to be defined and symbolically manipulated—through the institutionalized use of expert knowledge and political persuasion.

An expert approach for control of fermentation processes as variable structure plants

Structural variability is a fundamental property of biological systems. Fermentation processes, in particular, are often accompanied by physiological phenomena which have the characteristics of structural alterations. The conventional control approach does not have the competence to handle such processes, because it is based on the general assumption for a single-structure plant. In order to overcome this limitation, we consider under some conditions fermentation processes as variable structure plants and propose a two-level hierarchical scheme for their control. At the higher hierarchical level, which performs the organizing functions and operates in the structural space of the plant, the current plant structure is recognized as an element of a finite set of structures defined on the basis of ‘deep’ expert knowledge of the cell physiology and/or ‘shallow’ expert knowledge of the behavioral characteristics of the microbial population. The recognition is based on fuzzy decomposition of the structural space of the plant into several subspaces, mapped isomorphically into a respective set of rational control strategies. A set of expert rules is used for the actual synthesis of the recognition procedure, which is finally tuned by supervised learning. The adopted methodology at this hierarchical level provides the system with a kind of intelligence. To the lower hierarchical level, two distinct functions are assigned. The first one performs the global task of control of the plant structure and operates in its structural space. The second function has a local character and consists of control of the plant in a particular subspace of constant structure. It works in the state space of the plant. Under supervision of the higher level, the control strategy relevant to the current plant structure is picked out from a predefined pool and the corresponding control action is calculated. Extensive use of expert knowledge in the conceptual formulation of proper control strategies is assumed. Compared with the conventional single-structure approach, the variable structure concept results in rational and physiologically motivated control of fermentation processes. Instead of the development of a single but complicated process model and control structure, construction of several simple control structures, often in a ‘model-less’ form, is required. This method naturally incorporates and theoretizes the widely accepted policy of artificial induction of structural transformations in the plant, which is essential for enhanced productivity in many modern fermentation technologies. The proposed approach has been realized as a real-time software system, provided with enough flexibility to be used with different fermentation processes and equipment. This system is orientated to IBM personal computers and works under the QNX multi-tasking real-time operating system.

Decision making and classification

Work on the unification of methods of decision making and classification is reported here. Starting with decision at pixel level through region level, decisionsare taken at a planning level. Decision making at each level of extension has a common abstract structure. A major problem is the proper use of expert knowledge in automated decision making and classification. In artificial intelligence, the use of context in selecting decision rules is used. This approach is implemented by a technique of context masks which are first defined by the expert on a posterior basis. The context information is then used either on a-priori or posteriori basesin further, multitemporal data classification. Examples illustrate decision making up to land unit level and the use of expert knowledge in decision rules.