Systematic Literature Review on the symbolic knowledge extraction techniques

   page       attach   
Giovanni Mormone  •  Fabian Andres Aspee Encina

The development in the field of the AI, with the good performance of the so called sub-symbolic algorithms, has given the rise to the problem of the opacity of this methods, that take decisions based on their internal structure in an hard to understand way. To address this problem researchers proposed various symbolic knowledge extraction techniques, to overcome the opacity problem and extract human-readable rules. In this work we are going to perform a Systematic Literature Review (SLR) on the field of symbolic knowledge extraction to summarize and explain approaches, techniques, evaluation metrics and motivations behind knowledge extraction. To answer the research questions, we grouped together all the ambiguous terms found in the literature, to make the answers richer and to clarify the study field.