还原论抑或加工论:深度学习的根基反思与理论走向

Reductionism or Processing Theory:Fundamental Reflection and Theoretical Trend of Deep Learning

深度学习是学习者透过表象、超越知识、潜入心灵、全面领悟的一种知识学习现象、目标与状态,还原论与加工论是深度学习研究的两个理论原生点。问题、思维与意义是深度学习的三个关节点,回植情景、思维升级是深度学习创建的两大基本路径,贯通型深度学习理论是有机整合还原论与加工论的理念范型。面向未来,深度学习的理论走向是:构筑知识学习的立体世界,构建知识学习的生发链、生态环,建立学习力的系统提升机制。

Deep learning is a kind of knowledge learning phenomenon,goal and state that sees through the surface,transcends knowledge,penetrates into the heart and needs complete comprehension.Reductionism and processing theory are the two theoretical origins of deep learning research.The three joints of deep learning are problem,thinking and meaning,and the two basic paths of deep learning are scene-replanting and thinking upgrading.The penetrating deep learning theory is the conceptual paradigm of organic integration of reductionism and processing theory.In the future,the theoretical trend of deep learning is to construct a three-dimensional world of knowledge learning,build the generation chain and ecological ring of knowledge learning,and establish a systematic promotion mechanism of learning ability.