Transformer-based reranking for improving Korean morphological analysis systems
This study introduces a new approach in Korean morphological analysis combining dictionary-based techniques with Transformer-based deep learning models. The key innovation is the use of a BERT-based reranking system, significantly enhancing the accuracy of traditional morphological analysis. The met...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Electronics and Telecommunications Research Institute (ETRI)
2024-02-01
|
| Series: | ETRI Journal |
| Subjects: | |
| Online Access: | https://doi.org/10.4218/etrij.2023-0364 |
| _version_ | 1797266451679150080 |
|---|---|
| author | Jihee Ryu Soojong Lim Oh-Woog Kwon Seung-Hoon Na |
| author_facet | Jihee Ryu Soojong Lim Oh-Woog Kwon Seung-Hoon Na |
| author_sort | Jihee Ryu |
| collection | DOAJ |
| description | This study introduces a new approach in Korean morphological analysis combining dictionary-based techniques with Transformer-based deep learning models. The key innovation is the use of a BERT-based reranking system, significantly enhancing the accuracy of traditional morphological analysis. The method generates multiple suboptimal paths, then employs BERT models for reranking, leveraging their advanced language comprehension. Results show remarkable performance improvements, with the first-stage reranking achieving over 20% improvement in error reduction rate compared with existing models. The second stage, using another BERT variant, further increases this improvement to over 30%. This indicates a significant leap in accuracy, validating the effectiveness of merging dictionary-based analysis with contemporary deep learning. The study suggests future exploration in refined integrations of dictionary and deep learning methods as well as using probabilistic models for enhanced morphological analysis. This hybrid approach sets a new benchmark in the field and offers insights for similar challenges in language processing applications.
|
| first_indexed | 2024-04-25T01:00:54Z |
| format | Article |
| id | doaj.art-cbc3e98b1703412a8679ef5877ec41f1 |
| institution | Directory Open Access Journal |
| issn | 1225-6463 2233-7326 |
| language | English |
| last_indexed | 2024-04-25T01:00:54Z |
| publishDate | 2024-02-01 |
| publisher | Electronics and Telecommunications Research Institute (ETRI) |
| record_format | Article |
| series | ETRI Journal |
| spelling | doaj.art-cbc3e98b1703412a8679ef5877ec41f12024-03-11T04:13:48ZengElectronics and Telecommunications Research Institute (ETRI)ETRI Journal1225-64632233-73262024-02-0146113715310.4218/etrij.2023-0364Transformer-based reranking for improving Korean morphological analysis systemsJihee RyuSoojong LimOh-Woog KwonSeung-Hoon NaThis study introduces a new approach in Korean morphological analysis combining dictionary-based techniques with Transformer-based deep learning models. The key innovation is the use of a BERT-based reranking system, significantly enhancing the accuracy of traditional morphological analysis. The method generates multiple suboptimal paths, then employs BERT models for reranking, leveraging their advanced language comprehension. Results show remarkable performance improvements, with the first-stage reranking achieving over 20% improvement in error reduction rate compared with existing models. The second stage, using another BERT variant, further increases this improvement to over 30%. This indicates a significant leap in accuracy, validating the effectiveness of merging dictionary-based analysis with contemporary deep learning. The study suggests future exploration in refined integrations of dictionary and deep learning methods as well as using probabilistic models for enhanced morphological analysis. This hybrid approach sets a new benchmark in the field and offers insights for similar challenges in language processing applications. https://doi.org/10.4218/etrij.2023-0364deep learningkorean morphological analysisnatural language understandingpretrained transformer encoderreranking |
| spellingShingle | Jihee Ryu Soojong Lim Oh-Woog Kwon Seung-Hoon Na Transformer-based reranking for improving Korean morphological analysis systems ETRI Journal deep learning korean morphological analysis natural language understanding pretrained transformer encoder reranking |
| title | Transformer-based reranking for improving Korean morphological analysis systems |
| title_full | Transformer-based reranking for improving Korean morphological analysis systems |
| title_fullStr | Transformer-based reranking for improving Korean morphological analysis systems |
| title_full_unstemmed | Transformer-based reranking for improving Korean morphological analysis systems |
| title_short | Transformer-based reranking for improving Korean morphological analysis systems |
| title_sort | transformer based reranking for improving korean morphological analysis systems |
| topic | deep learning korean morphological analysis natural language understanding pretrained transformer encoder reranking |
| url | https://doi.org/10.4218/etrij.2023-0364 |
| work_keys_str_mv | AT jiheeryu transformerbasedrerankingforimprovingkoreanmorphologicalanalysissystems AT soojonglim transformerbasedrerankingforimprovingkoreanmorphologicalanalysissystems AT ohwoogkwon transformerbasedrerankingforimprovingkoreanmorphologicalanalysissystems AT seunghoonna transformerbasedrerankingforimprovingkoreanmorphologicalanalysissystems |