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

In Natural Language Processing (NLP), finding data augmentation techniquesthat can produce high-quality human-interpretable examples has always beenchallenging. Recently, leveraging kNN such that augmented examples areretrieved from large repositories of unlabelled sentences has made a steptoward interpretable augmentation. Inspired by this paradigm, we introduceMinimax-kNN, a sample efficient data augmentation strategy tailored forKnowledge Distillation (KD). We exploit a semi-supervised approach based on KDto train a model on augmented data. In contrast to existing kNN augmentationtechniques that blindly incorporate all samples, our method dynamically selectsa subset of augmented samples that maximizes KL-divergence between the teacherand student models. This step aims to extract the most efficient samples toensure our augmented data covers regions in the input space with maximum lossvalue. We evaluated our technique on several text classification tasks anddemonstrated that Minimax-kNN consistently outperforms strong baselines. Ourresults show that Minimax-kNN requires fewer augmented examples and lesscomputation to achieve superior performance over the state-of-the-art kNN-basedaugmentation techniques.

Not Far Away, Not So Close: Sample Efficient Nearest Neighbour Data Augmentation via MiniMax