How Far You Can Get Using Machine Learning Black-Boxes

Abstract

Abstract-Supervised Learning (SL) is a machine learning research area which aims at developing techniques able to take advantage from labeled training samples to make decisions over unseen examples. Recently, a lot of tools have been presented in order to perform machine learning in a more straightforward and transparent manner. However, one problem that is increasingly present in most of the SL problems being solved is that, sometimes, researchers do not completely understand what supervised learning is and, more often than not, publish results using machine learning black-boxes. In this paper, we shed light over the use of machine learning black-boxes and show researchers how far they can get using these out-of-thebox solutions instead of going deeper into the machinery of the classifiers. Here, we focus on one aspect of classifiers namely the way they compare examples in the feature space and show how a simple knowledge about the classifier's machinery can lift the results way beyond out-of-the-box machine learning solutions.

Towards an assessment grid for intelligent modeling assistance

“…As many heterogeneous modeling abstractions are used across the software life-cycle, modelers have to ensure the global consistency of these modeling abstractions and their relationships with other abstractions [39]. The recent use of black-box techniques, such as machine learning, also has an impact on understandability [19,33]. All the above have increased the demand for support for working with large models [10].…”

Opportunities in intelligent modeling assistance

“…The dichotomy and complementary nature of data and models are increasingly apparent in all phases of the software development life-cycle [28], i.e., not only during system use but also during system development, maintenance, and posterior evolution. Other contributing factors are (i) the development of modern datadriven software systems, which potentially involve different interdisciplinary perspectives [22,30], and (ii) the role of recent black-box techniques, such as machine learning, on understandability [36].…”