Concept Matching for Low-Resource Classification

Federico Errica, Ludovic Denoyer, Bora Edizel +4

We propose a model to tackle classification tasks in the presence of very little training data. To this aim, we approximate the notion of exact match with a theoretically sound mec…