How Far do We Get Using Machine Learning Black-Boxes?

With several good research groups actively working in machine learning (ML) approaches, we have now the concept of self-containing machine learning solutions that oftentimes work out-ofthe-box leading to the concept of ML black-boxes. Although it is important to have such blackboxes helping researchers to deal with several problems nowadays, it comes with an inherent problem increasingly more evident: we have observed that researchers and students are progressively relying on ML black-boxes and, usually, achieving results without knowing the machinery of the classi ̄ers. In this regard, this paper discusses the use of machine learning blackboxes and poses the question of how far we can get using these out-of-the-box solutions instead of going deeper into the machinery of the classi ̄ers. The paper focuses on three aspects of classi ̄ers: (1) the way they compare examples in the feature space; (2) the impact of using features with variable dimensionality; and (3) the impact of using binary classi ̄ers to solve a multi-class problem. We show how knowledge about the classi ̄er's machinery can improve the results way beyond out-of-the-box machine learning solutions.