Machine Learning

Machine learning methodologies, models, algorithms and techniques have become indispensable decision support tools in all industries, thanks to the ever-increasing availability of data. Machine learning models enable prediction, classification, anomaly identification, pattern recognition, time series analysis, in complex systems informally described by large amounts of data.

Deix is able to offer machine learning-based algorithmic solutions, thanks to a highly trained team and to the members of the Scientific Committee who have contributed to the development of state-of-the-art machine learning algorithms.

The particularity of Deix is that of data-driven optimization, i.e. of appropriately integrating machine learning and artificial intelligence techniques with optimization and control algorithms. The construction of the machine learning model obtained from the available data, which may be texts, images, videos, historical series is aimed at prediction, pattern recognition, classification, but it is also functional to the decision-making process above, which is guided by the optimization algorithm.

In supporting design and management decisions, Deix combines the typical tools of artificial intelligence (neural networks, convolutional networks, deep learning, clustering, feature selection, novelty detection) with optimization algorithms (constrained and unconstrained) and optimal control techniques.