@inproceedings{c8b15d69ef1c467a8eb68f1033718ea0,
title = "Extract Then Adjust: A Two-Stage Approach for Automatic Term Extraction",
abstract = "Automatic Term Extraction (ATE) is a fundamental natural language processing task that extracts relevant terms from domain-specific texts. Existing transformer-based approaches have indeed achieved impressive improvement. However, we observe that even state-of-the-art (SOTA) extractors suffer from boundary errors, which are distinguished by incorrect start or end positions of a candidate term. The minor differences between candidate terms and ground-truth leads to a noticeable performance decline. To alleviate the boundary errors, we propose a two-stage extraction approach. First, we design a span-based extractor to provide high-quality candidate terms. Subsequently, we adjust the boundaries of these candidate terms to enhance performance. Experiment results show that our approach effectively identifies and corrects boundary errors in candidate terms, thereby exceeding the performance of previous state-of-the-art models.",
keywords = "automatic term extraction, boundary adjust, span extraction",
author = "Jiangyu Wang and Chong Feng and Fang Liu and Xinyan Li and Xiaomei Wang",
note = "Publisher Copyright: {\textcopyright} The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.; 12th National CCF Conference on Natural Language Processing and Chinese Computing, NLPCC 2023 ; Conference date: 12-10-2023 Through 15-10-2023",
year = "2023",
doi = "10.1007/978-3-031-44696-2_19",
language = "English",
isbn = "9783031446955",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Science and Business Media Deutschland GmbH",
pages = "236--247",
editor = "Fei Liu and Nan Duan and Qingting Xu and Yu Hong",
booktitle = "Natural Language Processing and Chinese Computing - 12th National CCF Conference, NLPCC 2023, Proceedings",
address = "Germany",
}