Paragraph Boundary Recognition in Novels for Story Understanding
The understanding of narrative stories by computer is an important task for their automatic generation. To date, high-performance neural-network technologies such as BERT have been applied to tasks such as the Story Cloze Test and Story Completion. In this study, we focus on the text segmentation of...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-06-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/11/12/5632 |
| _version_ | 1797529679526100992 |
|---|---|
| author | Riku Iikura Makoto Okada Naoki Mori |
| author_facet | Riku Iikura Makoto Okada Naoki Mori |
| author_sort | Riku Iikura |
| collection | DOAJ |
| description | The understanding of narrative stories by computer is an important task for their automatic generation. To date, high-performance neural-network technologies such as BERT have been applied to tasks such as the Story Cloze Test and Story Completion. In this study, we focus on the text segmentation of novels into paragraphs, which is an important writing technique for readers to deepen their understanding of the texts. This type of segmentation, which we call “paragraph boundary recognition”, can be considered to be a binary classification problem in terms of the presence or absence of a boundary, such as a paragraph between target sentences. However, in this case, the data imbalance becomes a bottleneck because the number of paragraphs is generally smaller than the number of sentences. To deal with this problem, we introduced several cost-sensitive loss functions, namely. focal loss, dice loss, and anchor loss, which were robust for imbalanced classification in BERT. In addition, introducing the threshold-moving technique into the model was effective in estimating paragraph boundaries. As a result of the experiment on three newly created datasets, BERT with dice loss and threshold moving obtained a higher <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mi>F</mi><mn>1</mn></mrow></semantics></math></inline-formula> than the original BERT had using cross-entropy loss as its loss function (76% to 80%, 50% to 54%, 59% to 63%). |
| first_indexed | 2024-03-10T10:17:12Z |
| format | Article |
| id | doaj.art-0e5797fc59404588af49350884a2702a |
| institution | Directory Open Access Journal |
| issn | 2076-3417 |
| language | English |
| last_indexed | 2024-03-10T10:17:12Z |
| publishDate | 2021-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj.art-0e5797fc59404588af49350884a2702a2023-11-22T00:39:59ZengMDPI AGApplied Sciences2076-34172021-06-011112563210.3390/app11125632Paragraph Boundary Recognition in Novels for Story UnderstandingRiku Iikura0Makoto Okada1Naoki Mori2Graduate School of Engineering, Osaka Prefecture University, 1-1 Gakuen-cho, Naka-ku, Sakai, Osaka 599-8231, JapanGraduate School of Engineering, Osaka Prefecture University, 1-1 Gakuen-cho, Naka-ku, Sakai, Osaka 599-8231, JapanGraduate School of Engineering, Osaka Prefecture University, 1-1 Gakuen-cho, Naka-ku, Sakai, Osaka 599-8231, JapanThe understanding of narrative stories by computer is an important task for their automatic generation. To date, high-performance neural-network technologies such as BERT have been applied to tasks such as the Story Cloze Test and Story Completion. In this study, we focus on the text segmentation of novels into paragraphs, which is an important writing technique for readers to deepen their understanding of the texts. This type of segmentation, which we call “paragraph boundary recognition”, can be considered to be a binary classification problem in terms of the presence or absence of a boundary, such as a paragraph between target sentences. However, in this case, the data imbalance becomes a bottleneck because the number of paragraphs is generally smaller than the number of sentences. To deal with this problem, we introduced several cost-sensitive loss functions, namely. focal loss, dice loss, and anchor loss, which were robust for imbalanced classification in BERT. In addition, introducing the threshold-moving technique into the model was effective in estimating paragraph boundaries. As a result of the experiment on three newly created datasets, BERT with dice loss and threshold moving obtained a higher <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mi>F</mi><mn>1</mn></mrow></semantics></math></inline-formula> than the original BERT had using cross-entropy loss as its loss function (76% to 80%, 50% to 54%, 59% to 63%).https://www.mdpi.com/2076-3417/11/12/5632natural-language processingstory understandingtext segmentationimbalanced classificationBERTcost-sensitive loss |
| spellingShingle | Riku Iikura Makoto Okada Naoki Mori Paragraph Boundary Recognition in Novels for Story Understanding Applied Sciences natural-language processing story understanding text segmentation imbalanced classification BERT cost-sensitive loss |
| title | Paragraph Boundary Recognition in Novels for Story Understanding |
| title_full | Paragraph Boundary Recognition in Novels for Story Understanding |
| title_fullStr | Paragraph Boundary Recognition in Novels for Story Understanding |
| title_full_unstemmed | Paragraph Boundary Recognition in Novels for Story Understanding |
| title_short | Paragraph Boundary Recognition in Novels for Story Understanding |
| title_sort | paragraph boundary recognition in novels for story understanding |
| topic | natural-language processing story understanding text segmentation imbalanced classification BERT cost-sensitive loss |
| url | https://www.mdpi.com/2076-3417/11/12/5632 |
| work_keys_str_mv | AT rikuiikura paragraphboundaryrecognitioninnovelsforstoryunderstanding AT makotookada paragraphboundaryrecognitioninnovelsforstoryunderstanding AT naokimori paragraphboundaryrecognitioninnovelsforstoryunderstanding |