Automatic Dialect Adaptation in Finnish and its Effect on Perceived Creativity

TL;DR

This paper introduces a novel method for adapting standard Finnish to various dialects using character-level NMT models, and examines how dialectal variation influences perceived creativity in AI-generated poetry.

Contribution

It presents a transfer learning approach for dialect adaptation in Finnish and analyzes its impact on perceived creativity and originality.

Findings

Transfer learning slightly outperforms multi-dialectal models.

Greater dialect deviation reduces perceived creativity scores.

Dialectal fluency correlates with perceptions of originality.

Abstract

We present a novel approach for adapting text written in standard Finnish to different dialects. We experiment with character level NMT models both by using a multi-dialectal and transfer learning approaches. The models are tested with over 20 different dialects. The results seem to favor transfer learning, although not strongly over the multi-dialectal approach. We study the influence dialectal adaptation has on perceived creativity of computer generated poetry. Our results suggest that the more the dialect deviates from the standard Finnish, the lower scores people tend to give on an existing evaluation metric. However, on a word association test, people associate creativity and originality more with dialect and fluency more with standard Finnish.