Knowledge based

Knowledge-based action would include situations where the operator needs to contemplate the situation, interpret information or make a difficult decision. Also included in this grouping would be cases where a procedure is not well spelled out. In these cases the person performing the task must consider the actions to be taken and not act according to specific training. 

Knowledge-based errors are concerned with performance in novel or new situations. Actions have to be planned on-line and the process is intellectually demanding. The problem solver will only resort to this type of activity when he has run out of rule-based solutions. An example of knowledge-based performance is that of first learning to operate a piece of machinery. The hydraulic controls of a winch provide a good example. Experimentation will help the operator to build a mental model of how the controls can be co-ordinated to achieve the desired movements. Eventually, the operator will adopt a set of rules derived from that mental model. With practice, the task will become skill-based. Training offers the opportunity to miss out the experimentation phase by guiding the trainee to a correct model of situations, based on the experiences of others. 

One problem with the Rasmussen data is that it requires subjective analysis of the operator s training and capabilities. A set of human error rates was developed by Hunns for more specific tasks, not relying as much on the operator s capabilities and knowledge (Hunns (1982)). These data are presented in Table 9.3 and were based on extrapolation from human error rate databases. These data are similar to the rates of Rasmussen in Table 9.2 but provide some actual examples and do not require as much subjective analysis as the Rasmussen data. 

The human error rates for some specific tasks have been provided by Dhillon and are presented in Table 9.4 (Dhillon (1986)). Dhillon points out that there are six basic categories of error sources that can eventually lead to an accident condition  

Departure from following correct operating procedures. 

---

The classical computer tomography (CT), including the medical one, has already been demonstrated its efficiency in many practical applications. At the same time, the request of the all-round survey of the object, which is usually unattainable, makes it important to find alternative approaches with less rigid restrictions to the number of projections and accessible views for observation. In the last time, it was understood that one effective way to withstand the extreme lack of data is to introduce a priori knowledge based upon classical inverse theory (including Maximum Entropy Method (MEM)) of the solution of ill-posed problems [1-6]. As shown in [6] for objects with binary structure, the necessary number of projections to get the quality of image restoration compared to that of CT using multistep reconstruction (MSR) method did not exceed seven and eould be reduced even further. 

M.J. Sippl, Boltzmann s principle, knowledge based mean fields and protein folding, J. Comp. Aided Mol. Design 7 (1993), 473-501. 

The classical architecture of an expert system comprises a knowledge base, an inference engine, and some kind of user interface. Most expert systems also include an explanation subsystem and a knowledge acquisition subsystem. This architecture is given in Figure 9-34 and described in more detail below. 

Knowledge hose The knowledge base contains the encoded knowledge which is needed to solve a certain problem. The knowledge can be represented in the form of facts or rules. 

Knowledge acquisition subsystem The task of the knowledge acquisition subsystems is to assemble and upgrade the knowledge base. A major task is to verify the data and check for consistency. 

The EROS (Elaboration of Reactions for Organic Synthesis) system [26] is a knowledge-based system which was created for the simulation of organic reactions. Given a certain set of starting materials, EROS investigates the potential reaction pathways. It produces sequences of simultaneous and consecutive reactions and attempts to predict the products that will be obtained in those reactions. 

Correlations between structure and mass spectra were established on the basis of multivariate analysis of the spectra, database searching, or the development of knowledge-based systems, some including explicit management of chemical reactions. 

To become familiar with a knowledge-based reaction prediction system To appreciate the different levels in the evaluation of chemical reactions To know how reaction sequences are modeled To understand kinetic modeling of chemical reactions To become familiar with biochemical pathways... 

One of the first attempts to build a knowledge base for synthetic organic reactions was made by Gelernter s group, through inductive and deductive machine learning [1]. Important work on this topic was also performed by Funatsu and his group [2]. 

This reaction data set of 626 reactions was used as a training data set to produce a knowledge base. Before this data set is used as input to a neural Kohonen network, each reaction must be coded in the form of a vector characterizing the reaction event. Six physicochemical effects were calculated for each of five bonds at the reaction center of the starting materials by the PETRA (see Section 7.1.4) program system. As shown in Figure 10,3-3 with an example, the physicochemical effects of the two regioisomeric products arc different. 

The most recent version of EROS has a clearcut separation of the system proper, which performs all the manipulations on chemical structures and reactions, from the knowledge base, which defines the scope of it.s application (Figure 10.3-7). 

The knowledge base is essentially two-fold on one hand it consists of a series of procedures for calculating all-important physicochemical effects such as heats of reaction, bond dissociation energies, charge distribution, inductive, resonance, and polarizability effects (.see Section 7.1). The other part of the knowledge base defines the reaction types on which the EROS system can work. 

The Japanese program system AlPHOS is developed by Funatsu s group at Toyo-hashi Institute of Technology [40]. AlPHOS is an interactive system which performs the retrosynthetic analysis in a stepwise manner, determining at each step the synthesis precursors from the molecules of the preceding step. AlPHOS tries to combine the merits of a knowledge-based approach with those of a logic-centered approach. 

Ithough knowledge-based potentials are most popular, it is also possible to use other types potential function. Some of these are more firmly rooted in the fundamental physics of iteratomic interactions whereas others do not necessarily have any physical interpretation all but are able to discriminate the correct fold from decoy structures. These decoy ructures are generated so as to satisfy the basic principles of protein structure such as a ose-packed, hydrophobic core [Park and Levitt 1996]. The fold library is also clearly nportant in threading. For practical purposes the library should obviously not be too irge, but it should be as representative of the different protein folds as possible. To erive a fold database one would typically first use a relatively fast sequence comparison lethod in conjunction with cluster analysis to identify families of homologues, which are ssumed to have the same fold. A sequence identity threshold of about 30% is commonly... 

Barton G J 1998. Protein Sequence Aligrunent Techniques. Acta Crystallographica 054 1139-1146. Blundell T L, B L Sibanda, M J E Sterbnerg and J M Thornton. Knowledge-based Prediction of Prote Structures and the Design of Novel Molecules. Nature 326 347-352. 

M J1990. Calculation of Conformational Ensembles from Potentials of Mean Force. An Approach o the Knowledge-Based Prediction of Local Structures in Globular Proteins. Journal of Molecular Siology 213 859-883. 

Currunins D J, C W Andrews, J A Benfley and M Cory 1996. Molecular Diversity in Chemical Database Comparison of Medicinal Chemistry Knowledge Bases and Databases of Commercially Availabl Compounds Journal of Chemical Information and Computer Science 36 750-763. 

Ia the 1990s robotics guided by artificial iateUigeace are expected to play a role comparable to that of electronics in the 1940s and 1950s (6). Expert systems, which are knowledge-based systems that can effectively represent and apply factual knowledge in specific areas of human expertise, seem ideaUy suited to robot supervision. 

Computer-aided process synthesis systems do not mean completely automated design systems (57). Process synthesis should be carried out by interactive systems, in which the engineer s role is to carry out synthesis and the machine s role is to analy2e the performance of synthesized systems. Computet apphcations in the future will probably deal with the knowledge-based system in appHed artificial intelligence. Consequendy, research on computer-aided process synthesis should be directed toward the realization of such systems with the collaboration of experienced process engineers. 

---