Recent emergence of artificial creativity, consequent from spectacular latest advancements of artificial intelligence, questions traditional standpoints relating the “creativeness”, commonly considered as the exclusive expression of humans’ intellect. However, from our point of view, the purpose of artificial creativity cannot be addressed without considering the philosophical deliberations regarding the humans’ intellect and their self-sufficiency in aesthetic perception of the surrounding environment in which they evolve. In this paper, we present an Emotional Unsupervised Classifier (EUC) achieving the aesthetic artificial visual appreciation of visual information. The unsupervised nature of the proposed classifier makes the decision-making process an emergent process in opposition to the imitative nature of supervised classifiers. It is important for blending in aesthetic experience based appraisal. Operating as a self-organizing map, where the self-organization is issued from a hybrid dual process, its topological bearing follows an emotions model issued from behavioral psychology. This latest character of our EUC makes it explainable, thereby extending the interpretability of both its operational mechanism and the issued results. Experimental results validating the investigated EUC are reported and discussed.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

An Emotional Classifier for Machine’s Artificial Visual Aesthetic Appraisal

  • Fatemeh Saveh,
  • Mohand Tahar Soualah,
  • Kurosh Madani,
  • Abdennasser Chebira

摘要

Recent emergence of artificial creativity, consequent from spectacular latest advancements of artificial intelligence, questions traditional standpoints relating the “creativeness”, commonly considered as the exclusive expression of humans’ intellect. However, from our point of view, the purpose of artificial creativity cannot be addressed without considering the philosophical deliberations regarding the humans’ intellect and their self-sufficiency in aesthetic perception of the surrounding environment in which they evolve. In this paper, we present an Emotional Unsupervised Classifier (EUC) achieving the aesthetic artificial visual appreciation of visual information. The unsupervised nature of the proposed classifier makes the decision-making process an emergent process in opposition to the imitative nature of supervised classifiers. It is important for blending in aesthetic experience based appraisal. Operating as a self-organizing map, where the self-organization is issued from a hybrid dual process, its topological bearing follows an emotions model issued from behavioral psychology. This latest character of our EUC makes it explainable, thereby extending the interpretability of both its operational mechanism and the issued results. Experimental results validating the investigated EUC are reported and discussed.