【作 者】：
李小涛， 陈 川， 吴新全， 王海文
【关 键 词 】：
深度学习； 人工智能； 认识论； 深度认知
【栏 目】：
理论探讨
【中文摘要】：
人工智能研究方兴未艾，但其核心之一的深度学习却被教育领域误解、误用。为解决人工智能时代深度学习混淆视听的现状，努力从计算机科学、教育学、心理学、认识论哲学、逻辑学五个维度探讨深度学习，从深度学习的主体、世界权威期刊的讨论、浅层学习、算法概念、认识论哲学、认知心理学、逻辑学等方面澄清对深度学习的误解与误用。教育领域亟须认清这一现实：无论出于何种目的，计算机研究领域的深度学习绝不能混淆于教育领域的深度学习，当前对深度学习的误解，把计算机研究领域“算法对未知输出作出的预测”（机器领域的学习）误解为“人类的心理及行为的变化”（人类学习）来进行研究，严重违背科学求真求实的基本精神。为杜绝望文生义、混淆视听，教育领域应以深度认知（思维结果）来区别于计算机领域的深度学习（算法结果）。在人工智能时代，对深度学习及时进行澄清，有助于夯实理论基础，更好地服务于未来人工智能教育应用研究。
【英文摘要】：
The research of artificial intelligence is in the ascendant, but one of its core concepts "deep learning" has been misunderstood and misused in the field of education. In order to solve that problem, this paper tries to clarify the misunderstanding and misuse of deep learning from five dimensions of computer science, pedagogy, psychology, epistemology philosophy and logic, and from the subject of deep learning, the discussion of world leading journals, shallow learning, algorithm concept, epistemological philosophy, cognitive psychology and logic as well. In the field of education, it is urgent to recognize this reality: deep learning in the field of computer research must not be confused with that in the field of education. Currently, the "prediction of unknown output by algorithm" in the field of computer research (learning in the field of machine) is misunderstood as "the changes of human psychology and behaviors" (human learning), which is seriously inconsistent with the basic spirit of science to seek truth . In order to eliminate that confusion, deep cognition (thinking result) in the field of education should be distinguished from the deep learning(algorithm result) in the field of computer. In the era of AI, the timely clarification of deep learning will help to consolidate the theoretical foundation and better serve the application research of artificial intelligence education in the future.