Abstract:
A sentence is typically treated as the minimal syntactic unit used for extracting valuable information from a longer piece of text. However, in written Thai, there are no explicit sentence markers. We proposed a deep learning model for the task of sentence segmentation that includes three main contributions. First, we integrate n-gram embedding as a local representation to capture word groups near sentence boundaries. Second, to focus on the keywords of dependent clauses, we combine the model with a distant representation obtained from self-attention modules. Finally, due to the scarcity of labeled data, for which annotation is difficult and time-consuming, we also investigate and adapt two techniques, allowing us to utilize unlabeled data. The first one is Cross-View Training (CVT) as a semi-supervised learning technique, and the second one is a pre-trained language model (ELMo) to improve representation. In the Thai sentence segmentation experiments, our model reduced the relative error by 7.4% and 18.5% compared with the baseline models on the Orchid and UGWC datasets, respectively. We also applied our model to the task of punctuation restoration on the IWSLT English dataset. Our model outperformed the prior sequence tagging models, achieving a relative error reduction of 7.6%. Ablation studies revealed that utilizing n-gram representations was the main contributing factor for Thai, while the semi-supervised training helped the most for English.
