ProZe: Explainable and prompt-guided zero-shot text classification

Harrando, Ismail; Reboud, Alison; Schleider, Thomas; Ehrhart, Thibault; Troncy, Raphaël
IEEE Internet Computing, Special issue on knowledge-infused learning for computational social systems, 8 July 2022

As technology accelerates the generation and communication of textual data, the need
to automatically understand this content becomes a necessity. In order to classify text, being it for tagging, indexing or curating documents, one often relies on large, opaque models that are trained on pre-annotated datasets, making the process unexplainable, difficult to scale and ill-adapted for niche domains with scarce data. To tackle these challenges, we propose ProZe, a text classification approach that leverages knowledge from two sources: prompting pre-trained language models, as well as querying ConceptNet, a common-sense knowledge base which can be used to add a layer of explainability to the results. We evaluate our approach empirically and we show how this combination not only performs on par with state-of-the-art zero shot classification on several domains, but also offers explainable predictions that can be visualized.

DOI
HAL
Type:
Journal
Date:
2022-07-08
Department:
Data Science
Eurecom Ref:
6917
Copyright:
© 2022 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.

PERMALINK : https://www.eurecom.fr/publication/6917