Nacionalni portal odprte znanosti

Many machine learning problems can be interpreted as differing just in the level of supervision provided to the learning process. In this work we provide a unifying way of dealing with these different degrees of supervision. We show how the framework developed to accommodate this vision can deal wit ...