Implementation of eXplainable Artificial Intelligence Case Study on the Assessment of Movements to Support Neuromotor Rehabilitation

Solutions based on Artificial Intelligence are being used to solve problems in various domains. However, many people feel uncomfortable with this type of solution because they must understand how it works. In the face of this, the so-called eXplainable Artificial Intelligence arises, seeking not only to provide the answers produced by Artificial Intelligence but also to offer aspects of explainability, detailing the decision process and generating confidence. In this context, a literature review on eXplainable Artificial Intelligence has presented a brief comparative study between the most popular libraries for this implementation and a deepening of the theme of explainability evaluation and the comprehension process. A proposal for the implementation and evaluation of eXplainable Artificial Intelligence in the context of movement classification to support neuromotor rehabilitation was built from the results obtained. The first experiments performed showed to be promising. The proposal is expected to be relevant for addressing a growing theme in a context, the health area, that demands explicability and transparency in decisions.