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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

BERT-based models have shown a remarkable ability in the Chinese SpellingCheck (CSC) task recently. However, traditional BERT-based methods still sufferfrom two limitations. First, although previous works have identified thatexplicit prior knowledge like Part-Of-Speech (POS) tagging can benefit in theCSC task, they neglected the fact that spelling errors inherent in CSC data canlead to incorrect tags and therefore mislead models. Additionally, they ignoredthe correlation between the implicit hierarchical information encoded by BERT'sintermediate layers and different linguistic phenomena. This results insub-optimal accuracy. To alleviate the above two issues, we design aheterogeneous knowledge-infused framework to strengthen BERT-based CSC models.To incorporate explicit POS knowledge, we utilize an auxiliary task strategydriven by Gaussian mixture model. Meanwhile, to incorporate implicithierarchical linguistic knowledge within the encoder, we propose a novel formof n-gram-based layerwise self-attention to generate a multilayerrepresentation. Experimental results show that our proposed framework yields astable performance boost over four strong baseline models and outperforms theprevious state-of-the-art methods on two datasets.

Make BERT-based Chinese Spelling Check Model Enhanced by Layerwise Attention and Gaussian Mixture Model