PictoBERT: Transformers for next pictogram prediction

作者：

Highlights：

• AAC systems must help users find the most suitable pictograms to complete a phrase.

• Using Transformers for pictogram prediction may improve user communication.

• Our proposal outperformed the used in previous studies.

• The proposal’s main characteristic is the ability to transfer learning.

• The proposal allow fine-tuning to adapt to users’ needs.

摘要

•AAC systems must help users find the most suitable pictograms to complete a phrase.•Using Transformers for pictogram prediction may improve user communication.•Our proposal outperformed the used in previous studies.•The proposal’s main characteristic is the ability to transfer learning.•The proposal allow fine-tuning to adapt to users’ needs.

论文关键词：Augmentative and alternative communication,Language modeling,Pictogram prediction

论文评审过程：Received 22 June 2021, Revised 4 April 2022, Accepted 10 April 2022, Available online 20 April 2022, Version of Record 28 April 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.117231