Overfitting in quantum machine learning and entangling dropout

The ultimate goal in machine learning is to construct a model function that has generalization capability for unseen dataset, based on given training dataset. If the too much expressibility power, then it may overfit data and as result lose capability. To avoid such overfitting issue, several techniques have been developed classical regime, dropout one effective method. This paper proposes straightforward analogue of this technique quantum entangling dropout, meaning some gates parametrized circuit are randomly removed during process reduce circuit. Some simple case studies show actually suppresses overfitting.