Sm{\aa}prat: DialoGPT for Natural Language Generation of Swedish Dialogue by Transfer Learning

TL;DR

This paper explores adapting English transformer-based dialogue models to Swedish using transfer learning, demonstrating promising results in generating human-like responses with over 57% human-like scores.

Contribution

It presents the first empirical study of transfer learning for Swedish dialogue generation using DialoGPT, including fine-tuning methods and evaluation results.

Findings

Transfer learning effectively adapts DialoGPT to Swedish.

Over 57% of responses judged human-like by evaluators.

Models and demos are publicly available on HuggingFace.

Abstract

Building open-domain conversational systems (or chatbots) that produce convincing responses is a recognized challenge. Recent state-of-the-art (SoTA) transformer-based models for the generation of natural language dialogue have demonstrated impressive performance in simulating human-like, single-turn conversations in English. This work investigates, by an empirical study, the potential for transfer learning of such models to Swedish language. DialoGPT, an English language pre-trained model, is adapted by training on three different Swedish language conversational datasets obtained from publicly available sources. Perplexity score (an automated intrinsic language model metric) and surveys by human evaluation were used to assess the performances of the fine-tuned models, with results that indicate that the capacity for transfer learning can be exploited with considerable success. Human evaluators asked to score the simulated dialogue judged over 57% of the chatbot responses to be human-like for the model trained on the largest (Swedish) dataset. We provide the demos and model checkpoints of our English and Swedish chatbots on the HuggingFace platform for public use.