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/

Zendy Plus

Presents the access of premium content as premium feature

Premium Content

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

Zendy Tools

Zendy Open

While discriminative neural network classifiers are generally preferred,recent work has shown advantages of generative classifiers in term of dataefficiency and robustness. In this paper, we focus on natural languageinference (NLI). We propose GenNLI, a generative classifier for NLI tasks, andempirically characterize its performance by comparing it to five baselines,including discriminative models and large-scale pretrained languagerepresentation models like BERT. We explore training objectives fordiscriminative fine-tuning of our generative classifiers, showing improvementsover log loss fine-tuning from prior work . In particular, we find strongresults with a simple unbounded modification to log loss, which we call the"infinilog loss". Our experiments show that GenNLI outperforms bothdiscriminative and pretrained baselines across several challenging NLIexperimental settings, including small training sets, imbalanced labeldistributions, and label noise.

Discriminatively-Tuned Generative Classifiers for Robust Natural Language Inference