Electronic devices connected to the Internet (domestic or industrial), sometimes requires interaction with humans. At this point, the process of determining who access their resources, through the recognition of faces, voices, fingerprints, etc. and gestures of individuals is essential to enhance the experience of using them. This article describes the use of general cognitive model implemented in part by BoW (bag of words) and SVMs (support vector machines) used in artificial intelligence to support the automatic face recognition and learning.