AMALEA 2016 - International Workshop on Advances in Machine Learning

There are many real-world problems of such high complexity that traditional scientific approaches, based on physical and statistical modeling of the data generation mechanism, do not succeed in addressing them efficiently. The cause of inefficiencies often lies in multidimensionality, nonlinearities, chaotic phenomena and the presence of a plethora of degrees of freedom and unknown parameters in the mechanism that generates data. As a result, loss of information that is crucial to solve a problem is inherent in the data generation process itself, making a traditional mathematical solution intractable.