Still not there? Comparing Traditional Sequence-to-Sequence Models to Encoder-Decoder Neural Networks on Monotone String Translation Tasks
This provides practical guidance for researchers in digital humanities, text correction, and speech recognition on when to use specialized vs. generic methods for monotone translation tasks.
The paper compared encoder-decoder neural models against traditional methods on four monotone string translation tasks (OCR post-correction, spelling correction, grapheme-to-phoneme conversion, and lemmatization), finding that specialized traditional approaches often outperformed generic deep learning models.
We analyze the performance of encoder-decoder neural models and compare them with well-known established methods. The latter represent different classes of traditional approaches that are applied to the monotone sequence-to-sequence tasks OCR post-correction, spelling correction, grapheme-to-phoneme conversion, and lemmatization. Such tasks are of practical relevance for various higher-level research fields including digital humanities, automatic text correction, and speech recognition. We investigate how well generic deep-learning approaches adapt to these tasks, and how they perform in comparison with established and more specialized methods, including our own adaptation of pruned CRFs.