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

Eye movements in reading play a crucial role in psycholinguistic researchstudying the cognitive mechanisms underlying human language processing. Morerecently, the tight coupling between eye movements and cognition has also beenleveraged for language-related machine learning tasks such as theinterpretability, enhancement, and pre-training of language models, as well asthe inference of reader- and text-specific properties. However, scarcity of eyemovement data and its unavailability at application time poses a majorchallenge for this line of research. Initially, this problem was tackled byresorting to cognitive models for synthesizing eye movement data. However, forthe sole purpose of generating human-like scanpaths, purely data-drivenmachine-learning-based methods have proven to be more suitable. Followingrecent advances in adapting diffusion processes to discrete data, we proposeScanDL, a novel discrete sequence-to-sequence diffusion model that generatessynthetic scanpaths on texts. By leveraging pre-trained word representationsand jointly embedding both the stimulus text and the fixation sequence, ourmodel captures multi-modal interactions between the two inputs. We evaluateScanDL within- and across-dataset and demonstrate that it significantlyoutperforms state-of-the-art scanpath generation methods. Finally, we providean extensive psycholinguistic analysis that underlines the model's ability toexhibit human-like reading behavior. Our implementation is made available athttps://github.com/DiLi-Lab/ScanDL.

ScanDL: A Diffusion Model for Generating Synthetic Scanpaths on Texts