A Quantum Approach to Pattern Recognition and Machine Learning. Part II

Different classifier functions can be defined in the framework of a quantum approach to machine learning. While the fidelity-classifier is based on a measure of similarity between quantum states, other classifiers refer to the possibility of an empirical discrimination between different states. An important example is represented by the Helstrom-classifier that has been successfully applied to some empirical simulations, for instance to the study of bio-medical images. An interesting case is represented by the evaluation of clonogenic assays: a technique whose aim is measuring the survival-degree of in vitro-cell cultures, based on the ability of a single cell to grow and to form a colony. In this field a quantum approach allows us to increase the classification-accuracy, in comparison with the corresponding results that are currently obtained in the case of most classical approaches. Some open problems and some possible further developments are mentioned in the conclusion of the article.