Toward Zero-Shot and Zero-Resource Multilingual Question Answering
In recent years, multilingual question answering has been an emergent research topic and has attracted much attention. Although systems for English and other rich-resource languages that rely on various advanced deep learning-based techniques have been highly developed, most of them in low-resource...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2022-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9894421/ |
| _version_ | 1828388017973231616 |
|---|---|
| author | Chia-Chih Kuo Kuan-Yu Chen |
| author_facet | Chia-Chih Kuo Kuan-Yu Chen |
| author_sort | Chia-Chih Kuo |
| collection | DOAJ |
| description | In recent years, multilingual question answering has been an emergent research topic and has attracted much attention. Although systems for English and other rich-resource languages that rely on various advanced deep learning-based techniques have been highly developed, most of them in low-resource languages are impractical due to data insufficiency. Accordingly, many studies have attempted to improve the performance of low-resource languages in a zero-shot or few-shot manner based on multilingual bidirectional encoder representations from transformers (mBERT) by transferring knowledge learned from rich-resource languages to low-resource languages. Most methods require either a large amount of unlabeled data or a small set of labeled data for low-resource languages. In Wikipedia, 169 languages have less than 10,000 articles, and 48 languages have less than 1,000 articles. This reason motivates us to conduct a zero-shot multilingual question answering task under a zero-resource scenario. Thus, this study proposes a framework to fine-tune the original mBERT using data from rich-resource languages, and the resulting model can be used for low-resource languages in a zero-shot and zero-resource manner. Compared to several baseline systems, which require millions of unlabeled data for low-resource languages, the performance of our proposed framework is not only highly comparative but is also better for languages used in training. |
| first_indexed | 2024-12-10T06:05:13Z |
| format | Article |
| id | doaj.art-5d407e8aa7464daebc4608185d2cab6a |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-12-10T06:05:13Z |
| publishDate | 2022-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-5d407e8aa7464daebc4608185d2cab6a2022-12-22T01:59:43ZengIEEEIEEE Access2169-35362022-01-0110997549976110.1109/ACCESS.2022.32075699894421Toward Zero-Shot and Zero-Resource Multilingual Question AnsweringChia-Chih Kuo0Kuan-Yu Chen1https://orcid.org/0000-0001-9656-7551Computer Science and Information Engineering Department, National Taiwan University of Science and Technology, Taipei, TaiwanComputer Science and Information Engineering Department, National Taiwan University of Science and Technology, Taipei, TaiwanIn recent years, multilingual question answering has been an emergent research topic and has attracted much attention. Although systems for English and other rich-resource languages that rely on various advanced deep learning-based techniques have been highly developed, most of them in low-resource languages are impractical due to data insufficiency. Accordingly, many studies have attempted to improve the performance of low-resource languages in a zero-shot or few-shot manner based on multilingual bidirectional encoder representations from transformers (mBERT) by transferring knowledge learned from rich-resource languages to low-resource languages. Most methods require either a large amount of unlabeled data or a small set of labeled data for low-resource languages. In Wikipedia, 169 languages have less than 10,000 articles, and 48 languages have less than 1,000 articles. This reason motivates us to conduct a zero-shot multilingual question answering task under a zero-resource scenario. Thus, this study proposes a framework to fine-tune the original mBERT using data from rich-resource languages, and the resulting model can be used for low-resource languages in a zero-shot and zero-resource manner. Compared to several baseline systems, which require millions of unlabeled data for low-resource languages, the performance of our proposed framework is not only highly comparative but is also better for languages used in training.https://ieeexplore.ieee.org/document/9894421/Multilingual question answeringzero-shotzero-resourcemBERT |
| spellingShingle | Chia-Chih Kuo Kuan-Yu Chen Toward Zero-Shot and Zero-Resource Multilingual Question Answering IEEE Access Multilingual question answering zero-shot zero-resource mBERT |
| title | Toward Zero-Shot and Zero-Resource Multilingual Question Answering |
| title_full | Toward Zero-Shot and Zero-Resource Multilingual Question Answering |
| title_fullStr | Toward Zero-Shot and Zero-Resource Multilingual Question Answering |
| title_full_unstemmed | Toward Zero-Shot and Zero-Resource Multilingual Question Answering |
| title_short | Toward Zero-Shot and Zero-Resource Multilingual Question Answering |
| title_sort | toward zero shot and zero resource multilingual question answering |
| topic | Multilingual question answering zero-shot zero-resource mBERT |
| url | https://ieeexplore.ieee.org/document/9894421/ |
| work_keys_str_mv | AT chiachihkuo towardzeroshotandzeroresourcemultilingualquestionanswering AT kuanyuchen towardzeroshotandzeroresourcemultilingualquestionanswering |