An Evaluation of Neural Machine Translation Models on Historical Spelling Normalization

TL;DR

This paper evaluates various neural machine translation models for historical spelling normalization across five languages, demonstrating that NMT models outperform SMT, with specific architectures excelling under different data conditions.

Contribution

The study systematically compares multiple NMT architectures and attention mechanisms for spelling normalization, introducing a hybrid method that enhances performance.

Findings

NMT models outperform SMT in character error rate.

Transformer models need more data to outperform RNNs.

Subword models with small vocabularies are better for low-resource languages.

Abstract

In this paper, we apply different NMT models to the problem of historical spelling normalization for five languages: English, German, Hungarian, Icelandic, and Swedish. The NMT models are at different levels, have different attention mechanisms, and different neural network architectures. Our results show that NMT models are much better than SMT models in terms of character error rate. The vanilla RNNs are competitive to GRUs/LSTMs in historical spelling normalization. Transformer models perform better only when provided with more training data. We also find that subword-level models with a small subword vocabulary are better than character-level models for low-resource languages. In addition, we propose a hybrid method which further improves the performance of historical spelling normalization.