Showing papers on "Adaptive reasoning published in 1995"

Creative problem solving

Abstract: Problem solving, as commonly taught in schools, is an analytical or procedural approach. This approach almost exclusively employs left-brain thinking modes, is competitive, and relies on individual effort. However, creative problem solving is a framework that encourages whole-brain, iterative thinking in the most effective sequence; it is cooperative in nature and is most productive when done as a team effort.

Stratified case-based reasoning: reusing hierarchical problem solving episodes

Abstract: Stratified case-based reasoning is a technique in which abstract solutions produced during hierarchical problem solving are used to assist case-based retrieval, matching, and adaptation. We describe the motivation for the integration of case-based reasoning with hierarchical problem solving, exemplify its benefits, detail a set of algorithms that implement our approach, and present their comparative empirical evaluation on a path planning task. Our results show that stratified case-based reasoning significantly decreases the computational expense required to retrieve, match, and adapt cases, leading to performance superior both to simple case-based reasoning and to hierarchical problem solving ab initio.

Experiments On Adaptation-Guided Retrieval In Case-Based Design

Abstract: Case-based reasoning (CBR) has been applied with some success to complex planning and design tasks. In such systems, the best case is retrieved and adapted to solve a particular target problem. Often, the best case is that which can be most easily adapted to the target problem (as the overhead in adaptation is generally very high). Standard CBR systems use semantic-similarity to retrieve cases, on the assumption that the most similar case is the easiest case to adapt. However, this assumption can be shown to be flawed. In this paper, we report a novel retrieval method, called adaptation-guided retrieval, that is sensitive to the ease-of-adaptation of cases. In the context of a CBR system for software-design, called Deja Vu, we show through a series of experiments that adaptation-guided retrieval is more accurate than standard retrieval techniques, that it scales well to large case-bases and that it results in more efficient overall problem-solving performance. The implications of this method and these results are discussed.

Functional representation and reasoning for reflective systems

Abstract: Functional models have been extensively investigated in the context of several problemsolving tasks such as device diagnosis and design. In this paper, we view problem solvers themselves as devices, and use structure-behavior-function models to represent how they work. The model representing the functioning of a problem solver explicitly specifies how the knowledge and reasoning of the problem solver result in the achievement of its goals. Then, we employ these models for performance-driven reflective learning. We view performance-driven learning as the task of redesigning the knowledge and reasoning of the problem solver to improve its performance. We use the model of the problem solver to monitor its reasoning. Assign blame when it fails, and appropriately redesign its knowledge and reasoning. This paper focuses on the model-based redesign of a path planner's task structure. It illustrates the modelbased reflection using examples from an operational system called the Autognostic system.

Children's application of simultaneous and successive processing in inductive and deductive reasoning problems: Implications for developing scientific reasoning skills

Abstract: The research reported in this article was undertaken to obtain a better understanding of problem solving and scientific reasoning in 10-year-old children. The study involved measuring children's competence at syllogistic reasoning and in solving a series of problems requiring inductive reasoning. Children were also categorized on the basis of levels of simultaneous and successive synthesis. Simultaneous and successive synthesis represent two dimensions of information processing identified by Luria in a program of neuropsychological research. Simultaneous synthesis involves integration of information in a holistic or spatial fashion, whereas successive synthesis involves processing information sequentially with temporal links between stimuli. Analysis of the data generated in the study indicated that syllogistic reasoning and inductive reasoning were significantly correlated with both simultaneous and successive synthesis. However, the strongest correlation was found between simultaneous synthesis and inductive reasoning. These findings provide a basis for understanding the roles of spatial and verbal-logical ability as defined by Luria's neuropsychological theory in scientific problem solving. The results also highlight the need for teachers to provide experiences which are compatible with individual students' information processing styles.

Some considerations on Koczy's fuzzy interpolative reasoning method

Abstract: In the sparse fuzzy rule bases, conventional fuzzy reasoning methods encounter difficulty because of the lack of inference evidence. To tackle this problem, Koczy and Hirota have proposed a new fuzzy reasoning method called a linear interpolative reasoning method. In this paper, we analyze their reasoning method and find out that the reasoning consequences by the method become sometimes abnormal fuzzy sets. The reasoning conditions of the reasoning method also discussed analytically. >

Interval-valued fuzzy backward reasoning

Abstract: The importance and efficiency of backward reasoning in nonfuzzy reasoning has been stressed for a long time, especially in the case of expert systems and decision-support systems. The extension of this reasoning method to fuzzy theory, however, has never been considered. In this paper, the authors propose a definition of fuzzy backward reasoning based on the generalized modus ponens and show the necessity of considering interval-valued fuzzy backward reasoning. Then, the authors propose solving methods for fuzzy backward reasoning in the case of a rule with one or several conditions as well as in the case of several rules.

Learning to reason the non monotonic case

Abstract: We suggest a new approach for the study of the non monotonicity of human commonsense reasoning. The two main premises that underlie this work are that commonsense reasoning is an inductive phenomenon and that missing information in the interaction of the agent with the environment may be as informative for future interactions as observed information. This intuition is normalized and the problem of reasoning from incomplete information is presented as a problem of learning attribute functions over a generalized domain. We consider examples that illustrate various aspects of the non monotonic reasoning phenomena which have been used over the years as bench marks for various formalisms and translate them into Learning to Reason problems. We demonstrate that these have concise representations over the generalized domain and prove that these representations can be learned efficiently. The framework developed suggests an operational approach to studying reasoning that is nevertheless rigorous and amenable to analysis. We show that this approach efficiently supports reasoning with incomplete information and at the same lime matches our expectations of plausible patterns of reasoning in cases where other theories do not. This work continues previous works in the Learning to Reason framework and supports the thesis that in order to develop a computational account for commonsense reasoning one should study the phenomena of learning and reasoning together.

Qualitative structural analysis using diagrammatic reasoning

Abstract: Diagrammatic reasoning is a type of reasoning in which the primary means of inference is direct manipulation and inspection of a diagram. Diagrammatic reasoning is prevalent in human problem solving behavior, especially for problems involving spatial relations among physical objects. Our research examines the relationship between diagrammatic and symbolic reasoning in a computational framework. We have built a system, called REDRAW, that emulates the human capability for reasoning with pictures in civil engineering. The class of structural analysis problems chosen provides a realistic domain whose solution process requires domain-specific knowledge as well as pictorial reasoning skills. We hypothesize that diagrammatic representations provide an environment where inferences about the physical results of proposed structural configurations can take place in a more intuitive manner than that possible through purely symbolic representations.

A Comparitive Utility Analysis of Case-Based Reasoning and Control-Rule Learning Systems

Abstract: The utility problem in learning systems occurs when knowledge learned in an attempt to improve a system's performance degrades performance instead. We present a methodology for the analysis of utility problems which uses computational models of problem solving systems to isolate the root causes of a utility problem, to detect the threshold conditions under which the problem will arise, and to design strategies to eliminate it. We present models of case-based reasoning and control-rule learning systems and compare their performance with respect to the swamping utility problem. Our analysis suggests that case-based reasoning systems are more resistant to the utility problem than control-rule learning systems.

Modelling Medical Concepts as Time-Objects

Abstract: Time is intrinsically related to medical problem-solving in general Modelling time from the perspective of computer-based, competent, solution derivation of real-life medical problems is a challenging undertaking Starting from the premise that temporal reasoning is an integral aspect of medical problem-solving, necessary requirements for medical temporal reasoning are listed, the notion of a time-object is proposed as an appropriate ontological primitive for modelling medical concepts, and generic models for patient data, disorders, and actions based on time-objects are discussed

Towards the Integration of Case-Based, Schema-Based and Model-Based Reasoning for Supporting Complex Design Tasks

Abstract: This paper presents an approach of how to build bridges between case-based and model-based reasoning. Unlike other approaches, these bridges do not intend to surmount the whole ”abstraction distance” between concrete cases and generic models in one step. Instead they introduce and use a web of supporting columns which consist of useful intermediate representations called schemata, prototypes, patterns, templates or whatever is a suitable characterisation for their functional role in the design process.

Inter-diagrammatic reasoning

Abstract: Endowing a computer with an ability to reason with diagrams could be of great benefit in terms of both human-computer interaction and computational efficiency through explicit representation. To date, research in diagrammatic reasoning has dealt with intra-diagrammatic reasoning (reasoning with a single diagram) almost to the exclusion of inter-diagrammatic reasoning (reasoning with related groups of diagrams). We postulate a number of general inter-diagrammatic operators and show how such operators can be useful in various diagrammatic domains. We develop a heuristic in the domain of game notation, derive fingering information in the domain of musical notation, and infer new information from related cartograms.

An analysis of systematic approaches to reasoning about actions and change

Abstract: Systematic approaches to reasoning about dynamical systems and causality provide a new view on how to proceed toward a general and uniform semantical framework which is independent from specific solutions to the frame problem, say. This direction of research enables to rigorously comparing known methodologies designed for reasoning about actions and change with respect to such a semantics. Two actual systematic approaches, namely the Ego-World-Semantics and the Action Description Language , are analyzed and compared in this paper. We present two equivalence results for ontological subclasses and elaborate the major differences. The ultimate aim of this analysis shall be a proposal about a combination of the two frameworks to obtain a powerful semantics which profits from the merits of both the Ego-World-Semantics as well as the Action Description Language.

Default-reasoning with models

Abstract: Reasoning with model-based representations is an intuitive paradigm, which has been shown to be theoretically sound and to possess some computational advantages over reasoning with formula-based representations of knowledge. In this paper we present more evidence to the utility of such representations. In real life situations, one normally completes a lot of missing "context" information when answering queries. We model this situation by augmenting the available knowledge about the world with context-specific information; we show that reasoning with model-based representations can be done efficiently in the presence of varying context information. We then consider the task of default reasoning. We show that default reasoning is a generalization of reasoning within context, in which the reasoner has many "context" rules, which may be conflicting. We characterize the cases in which model-based reasoning supports efficient default reasoning and develop algorithms that handle efficiently fragments of Reiter's default logic. In particular, this includes cases in which performing the default reasoning task with the traditional, formula-based, representation is intractable. Further, we argue that these results support an incremental view of reasoning in a natural way.

The Road To Digitopolis: Perils of Problem Solving

Abstract: How students solve problems is a topic of central concern both to educational researchers and to math/science teachers: What is the nature of good and poor problem solving? How can students improve their problem-solving capacities? Teachers are in a unique position to witness problem solving in action, and to draw connections between the classroom experiences of their students and the findings of research. This article presents an instance of problem solving (drawn from a popular children's book) annotated with references to current research in cognition and education. The annotations explore issues such as the effect of performance anxiety on problem solving, how problem solvers handle the experience of confusion, and the role of self-monitoring and metacognition in problem solving.

Reasoning about Reasoning: A Course Project:

Abstract: A course project for a sophomore-/junior-level course in thinking, reasoning, and decision making is described. Students in the course observe and describe the reasoning style of a partner from the class on four different tasks (e.g., geometric analogies and moral dilemmas). Then, they compare the performance across the four tasks to develop an argument about how many distinct kinds of reasoning they have observed in their partner. The assignment is designed to illustrate course material on theories and models of reasoning in concrete ways. Another and more important objective is to elicit from students a degree of critical thinking and creativity as they develop their own models and conclusions.

Problem Solving with The Incredible Machine - An Experiment in Case-Based Reasoning

Abstract: Mehmet H. Göker, M.Sc.Eng., M.Sc. Prof. Dr.-Ing. Herbert Birkhofer Technische Hochschule Darmstadt Maschinenelemente und Konstruktionslehre Magdalenenstr. 4, D-64289 Darmstadt, Germany {goker, birkhofer}@muk.maschinenbau.th-darmstadt.de 1 Motivation and Goals In order to analyze the development, influence and application of experience during problem solving and to draw conclusions for the development of Case-Based Reasoning systems [Ko93, RiSch89] we observed and videotaped test persons while they were solving simple design problems with the computer program "The Incredible Machine". The video and audio recordings were protocolled and analyzed. 2 The Experiment 2.1 The Computer Program "The Incredible Machine" The computer program "The Incredible Machine" (TIM) simulates a design environment in which simple machines can be built by using the 45 provided elements. A selection of these elements is shown in Table 1. Figure 1 shows the main screen of the program. The machine is built and started in the main window. The parts can be selected from the list on the right hand side of the screen which can be scrolled using the arrows. By clicking on the field in the upper right hand corner of the screen, the environment (gravity, air pressure) is activated and the machine started. The point and bonus display at the bottom of the screen was of no interest to us and deactivated. 2.2 The Setup of the Experiment During the experiments the test persons were asked to build machines to solve the given assignments. To reduce the pressure they were told, that their machines will not be evaluated in any way and that their time is not limited. First the test persons had to read the introduction to the experiment and the computer program and were allowed to test the handling of the program. Then they had to solve four obligatory assignments in fixed order and to describe the elements they used in a description booklet. After the first four assignments, they were asked to select two more assignments out of the eight we provided, give the reason why they selected these particular assignments, and solve them. After all assignments were solved, the test persons had to write down the lessons they learned as if they wanted to give hints to a friend, taking this test the next day. Some of the test persons were also asked to solve the first assignment once more.

Practical Reasoning in Oscar

Abstract: It has long been my conviction that many of the problems encountered in artificial intelligence research are basically philosophical problems. In particular, in order to build an artificial rational agent, one must first have a clear theory of rationality to serve as a target for implementation. Accordingly, the OSCAR project is aimed at providing such a theory and building an AI system to implement it. In its present incarnation, OSCAR is a programmable architecture for a rational agent, based upon a general-purpose defeasible reasoner. To use this architecture to construct an actual agent, one must fill it out in various ways. This can be regarded as a matter of programming the architecture to implement proposed principles of rationality. For those who are skeptical about the very possibility of interesting AI systems, it is to be emphasized that this system is fully implemented, and available from the author for use by other researchers. I will begin by giving a very brief sketch of the general architecture, and then I will turn to some questions about practical reasoning that will constitute the main focus of this paper. These are questions that must be answered before OSCAR can become a full-fledged rational agent. On the conception of rationality embodied in OSCAR (discussed further in my [1993] and [1995]), a rational agent can be regarded as having four basic constituents: • One or more mechanisms for proposing goals. • A mechanism for evaluating the “goodness” of plans. • A mechanism for searching for and adopting plans on the basis of their comparative evaluations by the plan evaluator. • A mechanism for initiating action on the basis of adopted plans (together, possibly, with built-in or learned plan-schemas). These mechanisms constitute a system of practical cognition. On any theory of rationality, plan evaluation and adoption will be based in part on what beliefs the agent has about its situation. Accordingly, an important part of a rational agent is a system of epistemic cognition producing such beliefs. As will be seen below, in OSCAR, the bulk of the work involved in finding, evaluating, and choosing plans and directing action is done by epistemic cognition rather than by dedicated special-purpose modules devoted to practical cognition. The OSCAR architecture begins with a situation-evaluator, which produces a (real-measurable) degree of liking for the agent’s current situation. This is presumed to be sensitive to the agent’s beliefs about its situation. The likeability of a situation is the degree the agent would like it if it had true beliefs about all relevant aspects of the situation. The objective of the agent’s reasoning is to put itself in situations that are more likeable than the situations in which it would find itself if it did not take action. Ideally, plans are evaluated in terms of the expected likeability of their being adopted. This is just the mathematical expectation of the likeability of the situation-type consisting of their being adopted. Reasoning about expected likeabilities involves both reasoning about likeabilities and reasoning about probabilities. Such reasoning is computationally difficult, so the OSCAR

Formalizing Reasoning About Change: A Temporal Diagnosis Approach

Abstract: In this paper we describe a framework for reasoning about temporal explanation problems, which is based on our previous work on model-based diagnosis of dynamic systems. We use an explicit representation of qualitative temporal information which provides a simpler and more natural representation than the situation calculus. We show how to generate more specific explanations by instantiating explanations and assuming an Open World Assumption. We argue that a framework for reasoning about action should be able to deal with concurrent and durative actions and show how they can be represented in our system.

Reduction of competitive hypotheses by generalized conceptual knowledge in ES for production line design

Abstract: This paper describes some aspects of an expert system for assisting production line design. The expert system which uses hypothetical reasoning repeats same kinds of reasoning contents. In order to solve this problem, we develop the method to use inductive reasoning in the expert system. In the inductive reasoning, the method of generating the generalized conceptual knowledge for machine tools is proposed.

Fuzzy Control as Self-Organizing Adaptive Constraint-Oriented Problem Solving

Abstract: By introducing the notion of constraint-oriented fuzzy inference, we will show that it provides us ways of fuzzy control methods that has abilities of adaptation, learning and self-organization. The basic supporting techniques behind these abilities are “hard” processing by Artificial Intelligence or traditional computational framework and “soft” processing by Neural Network, Genetic Algorithm and Reinforcement Learning techniques. In the former processing, Qualitative Reasoning and Instance Generalization by Symbolic Reasoning play important role, while by the latter processing, fuzzy control becomes capable of learning, adaptation and evolutional self-organization.

Preface: The field of automated reasoning☆

Abstract: The term automated reasoning (first introduced in 1980) accurately describes the objective of the field, the automation of logical reasoning. This article introduces scientists to the field and then briefly describes the papers found in this special issue.