Still not there? Comparing Traditional Sequence-to-Sequence Models to Encoder-Decoder Neural Networks on Monotone String Translation Tasks

2016/10/25 by Schnober, Carsten, Eger, Steffen, Dinh, Erik-Lân Do +1
#Computation and Language (cs.CL) #FOS: Computer and information sciences

paper · doi:10.48550/arxiv.1610.07796

Abstract

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.

Related

• Is Encoder-Decoder Redundant for Neural Machine Translation?
• Deconvolution-Based Global Decoding for Neural Machine Translation
• Sequence-to-Sequence Models Can Directly Translate Foreign Speech
• Sequence-to-Sequence Neural Net Models for Grapheme-to-Phoneme Conversion
• Neural Machine Translation in Linear Time