Cognitive Science

Có nhiều bằng chứng cho thấy kiến thức chuyên môn dưới dạng các sơ đồ là yếu tố chính phân biệt các chuyên gia với những người mới trong kỹ năng giải quyết vấn đề. Bằng chứng cho thấy hoạt động giải quyết vấn đề truyền thống không hiệu quả trong việc tiếp thu sơ đồ cũng đang gia tăng. Người ta cho rằng một lý do chính cho sự không hiệu quả của giải quyết vấn đề như một công cụ học tập là do các quy trình nhận thức yêu cầu bởi hai hoạt động này không chồng chéo đủ và rằng việc giải quyết vấn đề truyền thống dưới dạng phân tích phương tiện - mục tiêu yêu cầu một khối lượng tương đối lớn khả năng xử lý nhận thức, do đó không còn khả dụng cho việc tiếp thu sơ đồ. Một mô hình tính toán và các bằng chứng thực nghiệm hỗ trợ cho lập luận này. Các hệ quả lý thuyết và thực tiễn cũng được thảo luận.

A theory of analogy must describe how the meaning of an analogy is derived from the meanings of its parts. In the structure‐mapping theory, the interpretation rules are characterized as implicit rules for mapping knowledge about a base domain into a target domain. Two important features of the theory are (a) the rules depend only on syntactic properties of the knowledge representation, and not on the specific content of the domains; and (b) the theoretical framework allows analogies to be distinguished cleanly from literal similarity statements, applications of abstractions, and other kinds of comparisons.

Two mapping principles are described: (a) Relations between objects, rather than attributes of objects, are mapped from base to target; and (b) The particular relations mapped are determined by systematicity, as defined by the existence of higher‐order relations.

Functional and mechanistic comparisons are made between several network models of cognitive processing: competitive learning, interactive activation, adaptive resonance, and back propagation. The starting point of this comparison is the article of Rumelhart and Zipser (1985) on feature discovery through competitive learning. All the models which Rumelhart and Zipser (1985) have described were shown in Grossberg (1976b) to exhibit a type of learning which is temporally unstable. Competitive learning mechanisms can be stabilized in response to an arbitrary input environment by being supplemented with mechanisms for learning top‐down expectancies, or templates; for matching bottom‐up input patterns with the top‐down expectancies; and for releasing orienting reactions in a mismatch situation, thereby updating short‐term memory and searching for another internal representation. Network architectures which embody all of these mechanisms were called adaptive resonance models by Grossberg (1976c). Self‐stabilizing learning models are candidates for use in real‐world applications where unpredictable changes can occur in complex input environments. Competitive learning postulates are inconsistent with the postulates of the interactive activation model of McClelland and Rumelhart (1981), and suggest different levels of processing and interaction rules for the analysis of word recognition. Adaptive resonance models use these alternative levels and interaction rules. The self‐organizing learning of an adaptive resonance model is compared and contrasted with the teacher‐directed learning of a back propagation model. A number of criteria for evaluating real‐time network models of cognitive processing are described and applied.

We present statistical analyses of the large‐scale structure of 3 types of semantic networks: word associations, WordNet, and Roget's Thesaurus. We show that they have a small‐world structure, characterized by sparse connectivity, short average path lengths between words, and strong local clustering. In addition, the distributions of the number of connections follow power laws that indicate a scale‐free pattern of connectivity, with most nodes having relatively few connections joined together through a small number of hubs with many connections. These regularities have also been found in certain other complex natural networks, such as the World Wide Web, but they are not consistent with many conventional models of semantic organization, based on inheritance hierarchies, arbitrarily structured networks, or high‐dimensional vector spaces. We propose that these structures reflect the mechanisms by which semantic networks grow. We describe a simple model for semantic growth, in which each new word or concept is connected to an existing network by differentiating the connectivity pattern of an existing node. This model generates appropriate small‐world statistics and power‐law connectivity distributions, and it also suggests one possible mechanistic basis for the effects of learning history variables (age of acquisition, usage frequency) on behavioral performance in semantic processing tasks.

A new diagnostic modeling system for automatically synthesizing a deep‐structure model of a student's misconceptions or bugs in his basic mathematical skills provides a mechanism for explaining why a student is making a mistake as opposed to simply identifying the mistake. This report is divided into four sections: The first provides examples of the problems that must be handled by a diagnostic model. It then introduces procedural networks as a general framework for representing the knowledge underlying a skill. The challenge in designing this representation is to find one that facilitates the discovery of misconceptions or bugs existing in a particular student's encoding of this knowledge. The second section discusses some of the pedagogical issues that have emerged from the use of diagnostic models within an instructional system. This discussion is framed in the context of a computer‐based tutoring/gaming system developed to teach students and student teachers how to diagnose bugs strategically as well as how to provide a better understanding of the underlying structure of arithmetic skills. The third section describes our uses of an executable network as a tool for automatically diagnosing student behavior, for automatically generating “diagnostic” tests, and for judging the diagnostic quality of a given exam. Included in this section is a discussion of the success of this system in diagnosing 1300 school students from a data base of 20.000 test items. The last section discusses future research directions.

Despite the accumulation of substantial cognitive science research relevant to education, there remains confusion and controversy in the application of research to educational practice. In support of a more systematic approach, we describe the Knowledge‐Learning‐Instruction (KLI) framework. KLI promotes the emergence of instructional principles of high potential for generality, while explicitly identifying constraints of and opportunities for detailed analysis of the knowledge students may acquire in courses. Drawing on research across domains of science, math, and language learning, we illustrate the analyses of knowledge, learning, and instructional events that the KLI framework affords. We present a set of three coordinated taxonomies of knowledge, learning, and instruction. For example, we identify three broad classes of learning events (LEs): (a) memory and fluency processes, (b) induction and refinement processes, and (c) understanding and sense‐making processes, and we show how these can lead to different knowledge changes and constraints on optimal instructional choices.

We review the major phenomena of skilled typing and propose a model for the control of the hands and fingers during typing. The model is based upon an Activation‐Trigger‐Schema system in which a hierarchical structure of schemata directs the selection of the letters to be typed and, then, controls the hand and finger movements by a cooperative, relaxation algorithm. The interactions of the patterns of activation and inhibition among the schemata determine the temporal ordering for launching the keystrokes. To account for the phenomena of doubling errors, the model has only “type” schemata—no “token” schemata—with only a weak binding between the special schema that signals a doubling, and its argument. The model exists as a working computer simulation and produces an output display of the hands and fingers moving over the keyboard. It reproduces some of the major phenomena of typing, including the interkeystroke interval times, the pattern of transposition errors found in skilled typists, and doubling errors. Although the model is clearly inadequate or wrong in some of its features and assumptions, it serves as a useful first approximation for the understanding of skilled typing.

The authors examined cultural preferences for formal versus intuitive reasoning among East Asian (Chinese and Korean), Asian American, and European American university students. We investigated categorization (Studies 1 and 2), conceptual structure (Study 3), and deductive reasoning (Studies 3 and 4). In each study a cognitive conflict was activated between formal and intuitive strategies of reasoning. European Americans, more than Chinese and Koreans, set aside intuition in favor of formal reasoning. Conversely, Chinese and Koreans relied on intuitive strategies more than European Americans. Asian Americans' reasoning was either identical to that of European Americans, or intermediate. Differences emerged against a background of similar reasoning tendencies across cultures in the absence of conflict between formal and intuitive strategies.

When learning a new math concept, should learners be first taught the concept and its associated procedures and then solve problems, or solve problems first even if it leads to failure and then be taught the concept and the procedures? Two randomized‐controlled studies found that both methods lead to high levels of procedural knowledge. However, students who engaged in problem solving before being taught demonstrated significantly greater conceptual understanding and ability to transfer to novel problems than those who were taught first. The second study further showed that when given an opportunity to learn from the failed problem‐solving attempts of their peers, students outperformed those who were taught first, but not those who engaged in problem solving first. Process findings showed that the number of student‐generated solutions significantly predicted learning outcomes. These results challenge the conventional practice of direct instruction to teach new math concepts and procedures, and propose the possibility of learning from one's own failed problem‐solving attempts or those of others before receiving instruction as alternatives for better math learning.