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

Nowadays, due to the breakthrough in natural language generation (NLG),including machine translation, document summarization, image captioning, etcNLG models have been encapsulated in cloud APIs to serve over half a billionpeople worldwide and process over one hundred billion word generations per day.Thus, NLG APIs have already become essential profitable services in manycommercial companies. Due to the substantial financial and intellectualinvestments, service providers adopt a pay-as-you-use policy to promotesustainable market growth. However, recent works have shown that cloudplatforms suffer from financial losses imposed by model extraction attacks,which aim to imitate the functionality and utility of the victim services, thusviolating the intellectual property (IP) of cloud APIs. This work targets atprotecting IP of NLG APIs by identifying the attackers who have utilizedwatermarked responses from the victim NLG APIs. However, most existingwatermarking techniques are not directly amenable for IP protection of NLGAPIs. To bridge this gap, we first present a novel watermarking method for textgeneration APIs by conducting lexical modification to the original outputs.Compared with the competitive baselines, our watermark approach achieves betteridentifiable performance in terms of p-value, with fewer semantic losses. Inaddition, our watermarks are more understandable and intuitive to humans thanthe baselines. Finally, the empirical studies show our approach is alsoapplicable to queries from different domains, and is effective on the attackertrained on a mixture of the corpus which includes less than 10\% watermarkedsamples.

Protecting Intellectual Property of Language Generation APIs with Lexical Watermark