Problem-solving Domains Research Articles

Building an Intelligent Tutoring System: Some Guidelines from a Study of Human Tutors

One reason that intelligent tutoring systems (ITSs) are rarely found outside of the research lab has to do with the guidelines available to developers of these systems. First, some of these guidelines are stated as general, abstract goals such as “adapt to the student.” What ITS developers need, however, are specific strategies and techniques which can be implemented in an ITS to accomplish those goals. Second, not all of the guidelines have an empirical basis. One solution to both of these problems is to study human tutors. This paper demonstrates this approach, and discusses an empirical study of human tutors which was conducted to address these issues. Specifically, it discusses 1) the knowledge acquisition method which we designed to capture the appropriate empirical data, 2) how we used this method to study human tutors and students in the medical problem-solving domain of immunohematology (blood banking), 3) several guidelines which appeared to drive the tutors' behavior (e.g., “limit the number of interrupts to the student”), and 4) specific tutoring strategies which can be incorporated into an ITS to make its behavior follow these guidelines.

Plans, goals and selection rules in the comprehension of computer programs

Abstract The notion of the programming plan has been proposed as a mechanism through which one can explain the nature of expertise in programming. Soloway and Ehrlich (1984) suggest that such expertise is characterized by the existence and use of programming plans. However, studies in other complex problem-solving domains, notably text editing, suggest that expertise is characterized not only by the possession of plan-related structures but also by the development of appropriate selection rules which govern the implementation of plans in appropriate situations (Card et al. 1980, Kay and Black 1984, 1986). This paper presents an experimental study which examines the role of programming plans in the context of skill development in programming. The results of this study suggest that plan-based structures cannot be used in isolation to explain novice/expert differences. Indeed, such structures appear to prevail at intermediate levels of skill. The major characteristic of expertise in programming would appear ...

The surgeon simulator

A new technique of small group surgical teaching has been developed wherein the surgeon takes on the role of the patient. This technique, which incorporates extensive and immediate formative evaluation, has all the advantages of simulation techniques while avoiding the major problems of training, scheduling and cost. This method has been used to teach a wide variety of surgical disease processes with the major emphasis being teaching patient management strategies. Sixty-one medical students have been taught using this method and they have found it superior to conventional seminar teaching, particularly in the domains of problem solving, patient management strategies and thought provocation.

Highs Are to Lows as Experts Are to Novices: Individual Differences in the Representation and Solution of Standardized Figural Analogies

Two experiments compared the strategies used by high and lower scorers on standardized figural analogy tests to represent and solve problems. In Experiment 1, subjects freely sorted completed analogy (A:a::B:b) problems into categories. High scorers categorized problems largely on the basis of well-constrained spatial transformations between problem terms; Low scorers sorted according to perceptual (i.e., shape) similarities among the figures constituting the problems. Experiment 2 compared the two groups' solution strategies in terms of specific patterns of eye movements used in viewing problem and answer terms. High scorers appeared to view problem terms in a more efficient fashion than did lower scorers. The lower scorers focused on noncritical figural relations (mapping figure A onto figure B), and they devoted more resources to processing the answer alternatives than did the high scorers. The pattern of results supports the view that high scorers tend to work in a more forward, or constructive, fashion, whereas lower scorers more often work backwards, using a response-elimination approach. Taken together, these findings converge to suggest specific aptitude-related differences in the representation and solution of standardized figural analogy problems. These differences bear some similarities to expert-novice differences in a number of other problem-solving domains.

CASE productivity perceptions of software engineering professionals

Computer-aided software engineering (CASE) is moving into the problem-solving domain of the systems analyst. The authors undertook a study to investigate the various functional and behavioral aspects of CASE and determine the impact it has over manual methods of software engineering productivity.

Interactive Elicitation and Exchange of Knowledge

Abstract This paper gives a state-of-the-art report on the use of techniques based on personal construct psychology to automate interactive elicitation of cognitive structures for group learning and decision making. It outlines the structure and key components of the RepGrid implementation on a network of Macintosh computers, which includes tools for interactive reperto y grid elicitation and analysis, and a system for intemctizely exchanging and comparing reperto y grids from multiple simultaneous users on the network. The users of RepGrid are attempting to extend and understand their own thinking and problem-solving capabilities by interacting w'th other people. They are able to: see the relationship of thrir points of view to those of others explore differing termin, logy for the same constructs become aware of differing constructs having the same terminology extend their own construct systems with those of others provide others with constructs they have found valuable explore a problem-solving domain ...

A Multivariate Analysis of Factors Differentiating Behaviorally and Emotionally Dysfunctional Children From Other Groups in School

We investigated the relative contribution of stressors, social support coping skills, and self-concept to understanding the social adjustment of various school populations. We divided 38 sixth-grade students into four different groups: emotionally disturbed (ED) students, "average" learners, students with developmental learning delays, and children who excelled academically. We used four different assessments to measure each of the four main constructs specifically. The results revealed that the four factors contributed to a meaningful differentiation of the emotionally and behaviorally disturbed group from the other groups. ED children were especially characterized by deficits in the social-cognitive problem-solving domain. The findings support the use of social-cognitive problem-solving abilities as an important mediator of social adjustment.

A therapy planning architecture that combines decision theory and artificial intelligence techniques

Through our experience with the ONCOCIN cancer therapy consultation system, we have identified a set of medical planning problems to which no single existing computer-based reasoning technique readily applies. In response to the need for automated assistance with this class of problems, we have devised a computer program called ONYX that combines decision-theoretic and artificial intelligence approaches to planning. We discuss our rationale for devising a new planning architecture and describe in detail how that architecture is implemented. The program's planning process consists of three steps: (i) the use of rules derived from therapy planning strategies to generate a small set of plausible plans, (ii) the use of knowledge about the structure and behavior of the human body to create simulations that predict possible consequences of each plan for the patient, and (iii) the use of decision theory to rank the plans according to how well the results of each simulation meet the treatment goals. This architecture explicitly manages the uncertainty inherent in many planning tasks, introduces a possible mechanism for the dissemination of decision-theoretic therapy advice, and potentially increases the number of problem solving domains in which expert system techniques can be effectively applied.

Experimental goal regression: A method for learning problem-solving heuristics

This research examines the process of learning problem solving with minimal requirements for a priori knowledge and teacher involvement. Experience indicates that knowledge about the problem solving task can be used to improve problem solving performance. This research addresses the issues of what knowledge is useful, how it is applied during problem solving, and how it can be acquired. For each operator used in the problem solving domain, knowledge is incrementally learned concerning why it is useful, when it is applicable, and what transformation it performs. The method of experimental goal regression is introduced for improving the learning rate by approximating the results of analytic learning. The ideas are formalized in an algorithm for learning and problem solving and demonstrated with examples from the domains of simultaneous linear equations and symbolic integration.

Experimental Goal Regression: A Method for Learning Problem-Solving Heuristics

This research examines the process of learning problem solving with minimal requirements for a priori knowledge and teacher involvement. Experience indicates that knowledge about the problem solving task can be used to improve problem solving performance. This research addresses the issues of what knowledge is useful, how it is applied during problem solving, and how it can be acquired. For each operator used in the problem solving domain, knowledge is incrementally learned concerning why it is useful, when it is applicable, and what transformation it performs. The method of experimental goal regression is introduced for improving the learning rate by approximating the results of analytic learning. The ideas are formalized in an algorithm for learning and problem solving and demonstrated with examples from the domains of simultaneous linear equations and symbolic integration.

Diagnosing Cognitive Errors: Statistical Pattern Classification Based on Item Response Theory

This paper introduces a probabilistic model that is capable of diagnosing and classifying cognitive errors in a general problem-solving domain. The model is different from the usual deterministic strategies common in the area of artificial intelligence because item response theory is utilized to handle the variability of response errors. As for illustrating the model, the dataset obtained from a 38-item fraction addition test is used, and the students’ responses are classified into 34 groups of misconceptions. These groups are predetermined by the result of an error analysis previously done, and validated with the error diagnostic program written by a typical formal logic approach.

A Gestalt perspective on computer simulations of cognitive processes

Impressive computer simulations of the thinking processes of novices confronting problems and powerful programs displaying artificial intelligence in the solution of problems in highly limited domains are among the recent achievements in the application of the modern computer to issues in cognitive psychology. Missing in such work is the crucial step of understanding, that is, grasping both what is crucial in any given problem and why it is crucial. Classical Gestalt analyses of productive thought emphasized precisely this phenomenon of insight. This article summarizes some of the Gestalt concepts and principles in the psychology of problem solving and creative thinking. Modern cognitive psychology has largely ignored, or has found intractable, the phenomenon of achieving an appropriate representation (an understanding) of a problem domain, and, until a more adequate theory of insight is developed, computer models of genuine human thinking and problem solving will remain incomplete. It is proposed that re-examination of the classical Gestalt literature on problem solving and productive thinking may help guide psychologists in fruitful directions in their effort to improve the simulation of complex human cognitive processes, particularly in the domain of problem solving.

When novices surpass experts: The difficulty of a task may increase with expertise.

A detailed look at the representations constructed by novice and expert computer programmers is presented. The issue is addressed by looking at the interaction between the representation naturally formed by programmers at each level of expertise and .an experimentally induced abstract or concrete mental set. The data suggest that, in the absence of an experimentally provided set, experts form abstract representations (denned here as what a program does), whereas novices form concrete representations (defined here as how a program functions). The data also suggest that appropriate sets can aid each group to form the representation not natural to them; however, these representations are not as stable as the preferred ones. The generality of the findings and the utility of the experts' representation of a task is discussed both in relation to computer programming and to problem solving in other domains. In the semantically rich domains of skilled problem solving (Bhaskar & Simon, 1977), the difference between experts and nonexperts is both qualitative and quantitative. Not only do experts perform better than novices on quantitative measures of skill but experimental manipulations also uncover qualitative differ- ences in the representations and strategies used by experts. Repeatedly, we find that the work- ing representations of experts are abstract conceptualizations of the original problem statement, whereas those of novices are less abstract and focus more on surface features of the problem. For example, in a recent ex- periment, Adelson (1981c) found that expert programmers used abstract, conceptually based representations when attempting to re- call programming material, whereas novices used syntactically based representations. Using a multitrial free-recall procedure, Adelson asked novice and expert programmers to recall a set of 16 lines of programming code that had been presented in random order. Although the subjects had not been told that the 16 lines

Child competence and psychiatric risk. II. Areas of relationship between child and family functioning.

Teacher and peer ratings of 65 children at risk to schizophrenia and other severe psychopathology were divided into three behavioral domains: problem solving, social-emotional, and compliance. Each child area was then related to variables generated from a family consensus Rorschach procedure in an effort to gain a greater understanding of the mechanism of risk involvement. Within the problem solving domain, parental communication deviance, nonacknowledgment of comments, and clear communication were the three family variables which consistently and linearly related to child competence at school. In the social-emotional area, only parental communication deviance displayed a linear relationship to the child's functioning at school. No significant findings occurred in the area of compliance behavior, and the roles of age of child and child IQ did not influence any of the findings. The data indicated that parental communication deviance may be an important variable in risk involvement.

Solving the problem of problem‐solving domains

Solving the problem of problem‐solving domains

Discovery-Oriented Behavior and Problem Solving.

Discovery-oriented behavior during the presolution stage of a problem-solving task was explored by videotaping and scoring the behavior of 60 male and female college students. They were presented with 14 different objects and told they would later have to suggest novel and useful combinations of three or more objects taken together. Behavioral variables that reflected touching and transformation of the objects were generated from the videotapes and correlated with number of solutions produced. Touching variables had high positive (significant) correlations with the criterion for females, but low positive (nonsignificant) correlations for males. Transformation variables had moderate negative correlations with the criterion for males (significant), but low negative (nonsignificant) correlations for females. Discovery-oriented behavior appears to be worthy of further study in the problem-solving domain.