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 Tools annual

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 monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Contrastive learning is widely used for sentence representation learning.Despite this prevalence, most studies have focused exclusively on English andfew concern domain adaptation for domain-specific downstream tasks, especiallyfor low-resource languages like Japanese, which are characterized byinsufficient target domain data and the lack of a proper training strategy. Toovercome this, we propose a novel Japanese sentence representation framework,JCSE (derived from ``Contrastive learning of Sentence Embeddings forJapanese''), that creates training data by generating sentences andsynthesizing them with sentences available in a target domain. Specifically, apre-trained data generator is finetuned to a target domain using our collectedcorpus. It is then used to generate contradictory sentence pairs that are usedin contrastive learning for adapting a Japanese language model to a specifictask in the target domain.  Another problem of Japanese sentence representation learning is thedifficulty of evaluating existing embedding methods due to the lack ofbenchmark datasets. Thus, we establish a comprehensive Japanese SemanticTextual Similarity (STS) benchmark on which various embedding models areevaluated. Based on this benchmark result, multiple embedding methods arechosen and compared with JCSE on two domain-specific tasks, STS in a clinicaldomain and information retrieval in an educational domain. The results showthat JCSE achieves significant performance improvement surpassing directtransfer and other training strategies. This empirically demonstrates JCSE'seffectiveness and practicability for downstream tasks of a low-resourcelanguage.

JCSE: Contrastive Learning of Japanese Sentence Embeddings and Its Applications