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

Controllable text generation (CTG) aims to generate text with desiredattributes, and decoding-time-based methods have shown promising performance onthis task. However, in this paper, we identify the phenomenon of AttributeCollapse for the first time. It causes the fluency of generated text to rapidlydecrease when the control strength exceeds a critical value, rendering the textcompletely unusable. This limitation hinders the effectiveness of decodingmethods in achieving high levels of controllability. To address this problem,we propose a novel lightweight decoding framework named Air-Decoding. Its mainidea is reconstructing the attribute distributions to balance the weightsbetween attribute words and non-attribute words to generate more fluent text.Specifically, we train prefixes by prefix-tuning to obtain attributedistributions. Then we design a novel attribute distribution reconstructionmethod to balance the obtained distributions and use the reconstructeddistributions to guide language models for generation, effectively avoiding theissue of Attribute Collapse. Experiments on multiple CTG tasks prove that ourmethod achieves a new state-of-the-art control performance.

Air-Decoding: Attribute Distribution Reconstruction for Decoding-Time Controllable Text Generation