Kenneth Kotovsky

Empirical tests of a theory of human acquisition of concepts for sequential patterns

Abstract The paper examines a body of empirical data on Ss performing the Thurstone Letter Series Completion task, in order to test the theory proposed by the authors in 1963 for explaining behavior on this task. The data confirm the theory in its main aspects, while indicating the need for some minor extensions and modifications. In particular the data show that subjects first discover the periodicity of the letter series, then construct a description of the pattern, and finally use the pattern description to make an extrapolation. Most of the pattern descriptions used by Ss fall within the pattern language defined in the earlier paper.

On the Benefits and Pitfalls of Analogies for Innovative Design: Ideation Performance Based on Analogical Distance, Commonness, and Modality of Examples

Drawing inspiration from examples by analogy can be a powerful tool for innovative design during conceptual ideation but also carries the risk of negative design outcomes (e.g., design fixation), depending on key properties of examples. Understanding these properties is critical for effectively harnessing the power of analogy. The current research explores how variations in analogical distance, commonness, and representation modality influence the effects of examples on conceptual ideation. Senior-level engineering students generated solution concepts for an engineering design problem with or without provided examples drawn from the U.S. Patent database. Examples were crossed by analogical distance (near-field vs. far-field), commonness (more vs. less-common), and modality (picture vs. text). A control group that received no examples was included for comparison. Effects were examined on a mixture of ideation process and product variables. Our results show positive effects of far-field and less-common examples on novelty and variability in quality of solution concepts. These effects are not modulated by modality. However, detailed analyses of process variables suggest divergent inspiration pathways for far-field vs. less-common examples. Additionally, the combination of far-field, less-common examples resulted in more novel concepts than in the control group. These findings suggest guidelines for the effective design and implementation of design-by-analogy methods, particularly a focus on far-field, less-common examples during the ideation process.

Agent-Based Synthesis of Electromechanical Design Configurations

A new automated approach to engineering design known as A-design is presented that creates design configurations through the interaction of software agents. By combining unique problem solving strategies, these agents are able to generate solutions to openended design problems. The A-design methodology makes several theoretical claims through its combination of multiagent systems, multiobjective design selection, and stochastic optimization, and is currently implemented to solve general electromechanical design problems. While this paper presents an overview of the theoretical basis for A-design, it primarily focuses on the method for representing electromechanical design configurations and the reasoning of the agents that construct these configurations. Results from an electromechanical test problem show the generality of the functional representation.

Implicit learning in problem solving : The role of working memory capacity

Participants solving the Balls and Boxes puzzle for the first time were slowed in proportion to the level of working memory (WM) reduction resulting from a concurrent secondary task. On a second and still challenging solution of the same puzzle, performance was greatly improved, and the same WM load did not impair problem-solving efficiency. Thus, the effect of WM capacity reduction was selective for the first solution of the puzzle, indicating that learning to solve the puzzle, a vital part of the first solution, is slowed by the secondary WM-loading task. Retrospective verbal reports, tests of specific puzzle knowledge, and a recognition test of potential strategies all indicated that participants were unaware of their knowledge of the puzzle, suggesting that it had been learned implicitly. Concurrent protocols collected from participants supported this conclusion and further suggested that participants were not aware of learning to solve the puzzle as this learning occurred. These results provide evidence that implicit learning depends on WM capacity and that implicit memory can play an important role in problem solving. Starting with the original conception of problem solving as search through a problem space presented by Newell and Simon (1972), progressively more complex models and theories of problem solving have provided increasingly detailed accounts of human problem-solving behavior. These theories of problem solving universally assume a long-term knowledge store of rules and problem-solving strategies and a separate working memory (WM) that provides limited processing space for the evaluation of strategies and planning of operator application (e.g., move) sequences. The limited capacity of WM makes it possible to investigate the interaction of memory and problem-solving processes through manipulation (e.g., reduction) of WM capacity. However, recent work on the neurological basis of memory has shown that the canonical separation of memory into two components (WM and long-term memory) significantly underestimates the complexity of human memory. A number of studies of the phenomena of implicit learning and preserved learning in patients with anterograde amnesia (cf. Seger, 1994, and Squire, Knowlton, & Musen, 1993, for reviews) have provided convergent evidence that there are several long-term memory stores dependent on different

The influence of open goals on the acquisition of problem-relevant information

There have been a number of recent findings indicating that unsolved problems, or open goals more generally, influence cognition even when the current task has no relation to the task in which the goal was originally set. It was hypothesized that open goals would influence what information entered the problem-solving process. Three studies were conducted to establish the effect of open goals on the acquisition of problem-relevant information. It was found that problem-relevant information, or hints, presented implicitly in a 2nd task in between attempts at solving problems aided problem solving. This effect cannot be attributed to strategic behavior after participants caught on to the manipulation, as most participants were not aware of the relationship. The implications of this research are discussed, including potential contributions to our understanding of insight, incubation, transfer, and creativity.

A-Design: Theory and Implementation of an Adaptive, Agent-Based Method of Conceptual Design

A new theory of engineering design known as A-Design is introduced that models the workings of realistic engineering design in a complex adaptive system of interacting software agents. The methodology is general enough to be used on a variety of search problems, however the motivation behind the research is to create design configurations. The system constructs meaningful designs from a catalog of electromechanical components based on a variety of user-defined objectives while accom—modating changes that might occur in the focus of the problem.

The Role of Functionality in the Mental Representations of Engineering Students: Some Differences in the Early Stages of Expertise

As engineers gain experience and become experts in their domain, the structure and content of their knowledge changes. Two studies are presented that examine differences in knowledge representation among freshman and senior engineering students. The first study examines recall of mechanical devices and chunking of components, and the second examines whether seniors represent devices in a more abstract functional manner than do freshmen. The most prominent differences between these 2 groups involve their representation of the functioning of groups of electromechanical components and how these groups of components interact to produce device behavior. Seniors are better able to construct coherent representations of devices by focusing on the function of sets of components in the device. The findings from these studies highlight some ways in which the structure and content of mental representations of design knowledge differ during the early stages of expertise acquisition.

Discovering Structure in Design Databases Through Functional and Surface Based Mapping

This work presents a methodology for discovering structure in design repository databases, toward the ultimate goal of stimulating designers through design-by-analogy. Using a Bayesian model combined with latent semantic analysis (LSA) for discovering structural form in data, an exploration of inherent structural forms, based on the content and similarity of design data, is undertaken to gain useful insights into the nature of the design space. In this work, the approach is applied to uncover structure in the U.S. patent database. More specifically, the functional content and surface content of the patents are processed and mapped separately, yielding structures that have the potential to develop a better understanding of the functional and surface similarity of patents. Structures created with this methodology yield spaces of patents that are meaningfully arranged into labeled clusters, and labeled regions, based on their functional similarity or surface content similarity. Examples show that cross-domain associations and transfer of knowledge based on functional similarity can be extracted from the function based structures, and even from the surface content based structures as well. The comparison of different structural form types is shown to yield different insights into the arrangement of the space, the interrelationships between the patents, and the information within the patents that is attended to—enabling multiple representations of the same space to be easily accessible for design inspiration purposes. In addition, the placement of a design problem in the space effectively points to the most relevant cluster of patents in the space as an effective starting point of stimulation. These results provide a basis for automated discovery of cross-domain analogy, among other implications for creating a computational design stimulation tool.

Simulated Annealing and the Generation of the Objective Function: A Model of Learning During Problem Solving

A computational model of problem solving based on significant aspects of human problem solving is introduced. It is observed that during problem solving humans often start searching more or less randomly, becoming more deterministic over time as they learn more about the problem. This two‐phase aspect of problem‐solving behavior and its relation to learning is one of the important features this model accounts for. The model uses an accelerated simulated annealing technique as a search mechanism within a real‐time dynamic programming‐like framework upon a connected graph of neighboring problem states. The objective value of each node is adjusted as the model moves between nodes, learning more accurate values for the nodes and also compensating for misleading heuristic information as it does so. In this manner the model is shown to learn to more effectively solve isomorphs of the Balls and Boxes and Tower of Hanoi problems. The major issues investigated with the model are (a) whether such a simulated annealing‐based model exhibits the kind of random‐to‐directed transition in behavior exhibited by people, and (b) whether the progressive discovery of the objective function, even when given very little or poor initial information, is a plausible method for representing the learning that occurs during problem solving and the knowledge that results from that learning.

The Effect of Incidental Hints When Problems Are Suspended Before, During, or After an Impasse

Two studies examine how the time at which problem solving is suspended relative to an impasse affects the impact of incidental hints. An impasse is a point in problem solving at which a problem solver is not making progress and does not know how to proceed. In both studies, work on remote associates problems was suspended before an impasse was reached, at the time an impasse was reached, or after a period of continued work during an impasse. After problem solving was suspended on a set of problems, participants completed a lexical decision task before resuming work on the set of unsolved problems. For half of the problems suspended during each impasse state, solution words were presented as incidental hints in the lexical decision task. The proportion of initially unsolved problems that were solved after the intervening lexical decision task was greater when problem solving was suspended at the point an impasse was reached than when problem solving was suspended before an impasse or after a period of continued work during an impasse. These results suggest that suspending problem solving at the point of impasse may increase susceptibility to incidentally presented hints. The point of impasse may be an opportune time for hints because the problem has been explored but there has not been a large increase in fixation on failed solution attempts.