Author ORCID Identifier

0000-0003-2344-7377

Document Type

Other

Disciplines

Computer Sciences, Information Science, *human – machine relations

Publication Details

30th Irish Conference on Artificial Intelligence and Cognitive Science. AICS 2022. Munster, Ireland, December 8–9, 2022. Extended Abstracts.

chrome-extension://efaidnbmnnnibpcajpcglclefindmkaj/https://link.springer.com/content/pdf/10.1007/978-3-031-26438-2.pdf?pdf=button

Abstract

The style of generated text is how something is said rather than what is said. We hypothesize that changing the style of generated text can change the perceived personality of the text generation agent. Dialogue systems that aim to imitate a human agent can appear to have a consistent personality through a consistent, controllable style of conversation. Some recent work on the style of generated text [1] performs impressively in the small number of domains selected for their experiments using transformer and LSTM-based models. Lin et al. [1] used weak supervised learning as their data set lacks parallel data. The model reconstructs a given sentence using the style of another sentence in the same domain. While our work is still in progress, we see the principle of automatically learning styles from text types and applying them to new information sources as having great application potential in the long run. However, to test the scalability and generalisability of these methods we selected a large, multi-domain data set, ToTTo [2]. ToTTo is comprised of Wikipedia tables paired with descriptions and has 120,000 training examples. We transformed the data set into a format suitable to retrain Lin et al.’s [1] model using automated extraction techniques.

Funder

Science Foundation Ireland

Creative Commons License

Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License
This work is licensed under a Creative Commons Attribution-NonCommercial-Share Alike 4.0 International License.

Download
Contact the Author


Share

COinS