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

Dialect identification (DID) is a special case of general languageidentification (LID), but a more challenging problem due to the linguisticsimilarity between dialects. In this paper, we propose an end-to-end DID systemand a Siamese neural network to extract language embeddings. We use bothacoustic and linguistic features for the DID task on the Arabic dialectalspeech dataset: Multi-Genre Broadcast 3 (MGB-3). The end-to-end DID system wastrained using three kinds of acoustic features: Mel-Frequency CepstralCoefficients (MFCCs), log Mel-scale Filter Bank energies (FBANK) andspectrogram energies. We also investigated a dataset augmentation approach toachieve robust performance with limited data resources. Our linguistic featureresearch focused on learning similarities and dissimilarities between dialectsusing the Siamese network, so that we can reduce feature dimensionality as wellas improve DID performance. The best system using a single feature set achieves73% accuracy, while a fusion system using multiple features yields 78% on theMGB-3 dialect test set consisting of 5 dialects. The experimental resultsindicate that FBANK features achieve slightly better results than MFCCs.Dataset augmentation via speed perturbation appears to add significantrobustness to the system. Although the Siamese network with language embeddingsdid not achieve as good a result as the end-to-end DID system, the twoapproaches had good synergy when combined together in a fused system.

Convolutional Neural Networks and Language Embeddings for End-to-End Dialect Recognition