关键词: 运动专家知识系统, 表征学习, 自动性, 抽象化, 生成模型

Abstract:

At present, the mechanism by which athletic sports expert acquires and transfers, refines and updates their professional knowledge and skills through long-term training has not been clearly elaborated. Based on the expert-novice paradigm, the cognitive superiority of athletic sports expert from the perspective of representation learning is mainly embodied in attention superiority and memory superiority, and the key reasons for the cognitive superiority are the automaticity and abstraction of the knowledge and skills of athletic sports expert. Although the knowledge superiority of competitive sports expert is manifested in both procedural knowledge and declarative knowledge, skill learning is more manifested in the automatic extraction of procedural memory. The characteristics of automaticity and deliberate practice partly explain how athletes extract and update professional knowledge and skills from long-term training, while the abstract representation helps athletes acquire and transfer professional knowledge and skills. Moreover, the abstract representation is not only reflected in the schema stored in the expert’s long-term memory system, but also shows that the expert pay more attention to the overall characteristics of the stimulus information, which has the overall perceptual superiority, such as the memorization of the board by the Go master and the recognition of the stance pattern by the football player.

Then, from the perspective of generative model, the abstract representation mechanism of competitive sports expert’s knowledge and skills is discussed. In particular, the latest research results of Bayesian cognitive model and deep generative model are used for reference, and the mechanism of competitive sports expert’s acquisition and transfer of professional knowledge and skills from long-term training is deeply discussed, so as to construct a more complete competitive sports expert knowledge system. Research based on Bayesian cognitive model shows that the information contained in the motor schema of the expert is more abstract than that of the novice, which is speculated to be because the abstract task representation is conducive to the acquisition and transfer of professional knowledge and skills. The core motor schema stored in the expert memory is reconstructed by deliberate practice under different task rules in order to better match the characteristics of the task, which may help athletes to form high-level special skills closely related to the task rules. As a result, the abstraction of knowledge representation mechanism of athletic sports expert is discussed from the perspective of generative model, which provides a new theoretical basis for understanding the cognitive superiority of athletic sports expert and helping them break through the cognitive limitations.

It is interesting that, from the perspective of feature representation theory, athletes compress the information in similar task situations through long-term training and competition, extract semantic features to construct core motor schema and store them in long-term memory, and then generate specific motor program based on core motor schema in new tasks. From the perspective of prototype theory, if the expected goal is not achieved, the athlete then compares the executed exercise program with the best action example representation in memory according to the feedback execution error, and improves the motor program.

In recent years, with the wide application of neuroscience technology, such as EEG, near infrared, functional magnetic resonance imaging, etc., the research on the cognitive superiority of competitive sports expert has accumulated a lot of cognitive neural evidence, and researchers have also made a lot of active exploration in the embodied cognitive view of motor skill acquisition. To sum up, the existing studies have explored the automaticity characteristics and abstract representation of the knowledge and skills of competitive sports expert from the aspects of cognitive neural evidence and embodied cognition theory. In the future, sports psychologists should make more use of the technical means of neuroscience and the theoretical basis of cognitive psychology to understand the formation mechanism of cognitive superiority of competitive sports expert more deeply and comprehensively.

Key words: knowledge-based system of sports expert, representational learning, automaticity, abstraction, generative model