English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Current state-of-the-art approaches to text classification typically leverageBERT-style Transformer models with a softmax classifier, jointly fine-tuned topredict class labels of a target task. In this paper, we instead propose analternative training objective in which we learn task-specific embeddings oftext: our proposed objective learns embeddings such that all texts that sharethe same target class label should be close together in the embedding space,while all others should be far apart. This allows us to replace the softmaxclassifier with a more interpretable k-nearest-neighbor classificationapproach. In a series of experiments, we show that this yields a number ofinteresting benefits: (1) The resulting order induced by distances in theembedding space can be used to directly explain classification decisions. (2)This facilitates qualitative inspection of the training data, helping us tobetter understand the problem space and identify labelling quality issues. (3)The learned distances to some degree generalize to unseen classes, allowing usto incrementally add new classes without retraining the model. We presentextensive experiments which show that the benefits of ante-hoc explainabilityand incremental learning come at no cost in overall classification accuracy,thus pointing to practical applicability of our proposed approach.

Task-Specific Embeddings for Ante-Hoc Explainable Text Classification