Understanding Deep Learning by Methodology in the Dialectics of Nature
- https://doi.org/10.2991/assehr.k.201215.385How to use a DOI?
- AI, Deep Learning, Methodology
In this paper, deep learning is discussed which is considered a branch in the modern AI subject with fastest development. First of all, the history of the deep learning is reviewed and then the nature of the black-box in the deep learning is pointed out. With the nature of the black-box (a device, system or object which can be viewed in terms of its inputs and outputs, without any knowledge of its internal workings) in the deep learning at center, from the view of the methodology in the dialectics of nature, two aspects, repetitiveness and interpretability, are analyzed in the paper. In the modern deep learning theory system, repetitiveness and interpretability are most needed to be explored and complemented. On this basis, the insufficiency of the spirit of seeking truth from facts is discussed across the whole practical process from subject selection to application in the deep learning. Finally, based on the above discussions, it can be concluded that though facing many shortcomings with regard to methodology, the above mentioned obstacles can be removed with help of the deep learning, thus contributing greatly to the development of the AI which is always the constant objective of human by unremitting efforts over time.
- © 2020, the Authors. Published by Atlantis Press.
- Open Access
- This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).
Cite this article
TY - CONF AU - Zirui Cao AU - Yanghao Li PY - 2020 DA - 2020/12/17 TI - Understanding Deep Learning by Methodology in the Dialectics of Nature BT - Proceedings of the 2nd International Conference on Literature, Art and Human Development (ICLAHD 2020) PB - Atlantis Press SP - 22 EP - 26 SN - 2352-5398 UR - https://doi.org/10.2991/assehr.k.201215.385 DO - https://doi.org/10.2991/assehr.k.201215.385 ID - Cao2020 ER -