Ask2Transformers: Zero-Shot Domain labelling with Pre-trained Language Models

In this paper we present a system that exploits different pre-trained\nLanguage Models for assigning domain labels to WordNet synsets without any kind\nof supervision. Furthermore, the system is not restricted to use a particular\nset of domain labels. We exploit the knowledge encoded within different\noff-the-shelf pre-trained Language Models and task formulations to infer the\ndomain label of a particular WordNet definition. The proposed zero-shot system\nachieves a new state-of-the-art on the English dataset used in the evaluation.\n