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 Plus annual

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

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Self-supervised pre-training of transformer models has shown enormous successin improving performance on a number of downstream tasks. However, fine-tuningon a new task still requires large amounts of task-specific labelled data toachieve good performance. We consider this problem of learning to generalize tonew tasks with few examples as a meta-learning problem. While meta-learning hasshown tremendous progress in recent years, its application is still limited tosimulated problems or problems with limited diversity across tasks. We developa novel method, LEOPARD, which enables optimization-based meta-learning acrosstasks with different number of classes, and evaluate different methods ongeneralization to diverse NLP classification tasks. LEOPARD is trained with thestate-of-the-art transformer architecture and shows better generalization totasks not seen at all during training, with as few as 4 examples per label.Across 17 NLP tasks, including diverse domains of entity typing, naturallanguage inference, sentiment analysis, and several other text classificationtasks, we show that LEOPARD learns better initial parameters for few-shotlearning than self-supervised pre-training or multi-task training,outperforming many strong baselines, for example, yielding 14.5% averagerelative gain in accuracy on unseen tasks with only 4 examples per label.

Learning to Few-Shot Learn Across Diverse Natural Language Classification Tasks