MERAL Myanmar Education Research and Learning Portal
Item
{"_buckets": {"deposit": "cef27ac2-0593-41f7-ba09-29c0d721aeda"}, "_deposit": {"created_by": 73, "id": "7672", "owner": "73", "owners": [73], "owners_ext": {"displayname": "", "username": ""}, "pid": {"revision_id": 0, "type": "depid", "value": "7672"}, "status": "published"}, "_oai": {"id": "oai:meral.edu.mm:recid/00007672", "sets": ["user-miit"]}, "communities": ["miit"], "item_1583103067471": {"attribute_name": "Title", "attribute_value_mlt": [{"subitem_1551255647225": "Statistical Machine Translation between Myanmar Sign Language and Myanmar SignWriting", "subitem_1551255648112": "en"}]}, "item_1583103085720": {"attribute_name": "Description", "attribute_value_mlt": [{"interim": "This paper contributes the first evaluation of automatic machine translation between Myanmar Sign Language (MSL) and Myanmar SignWriting (MSW). The main motivation is to introduce SignWriting\nto the Myanmar Deaf society with the help of statistical machine translation. In this paper, we use our MSL-MSW corpus for general domain that contains a textual representation of MSL and its equivalent Myanmar\nSignWriting. The methods studied in this work were phrase-based, hierarchical phrase-based and the operation sequence model. In addition, two different segmentation schemes were studies, these were syllable segmentation and word segmentation for MSL. The\nperformance of the machine translation systems was automatically measured in terms of BLEU and RIBES for all experiments. Our main findings were that operation sequence model gave the highest scores (37.54 BLEU and 0.8280 RIBES) for MSL to MSW translation and hierarchical phrase based machine translation gave the highest scores (52.79 BLEU and 0.8756 RIBES) for MSW to MSL translation. Generally, translation with word segmented MSL achieved better performance\nthan syllable segmentation of MSL. Our 10-fold cross validation results produced promising results even with the limited training data and we expect this can be developed into a useful machine translation system as\nmore data becomes available in the future."}]}, "item_1583103108160": {"attribute_name": "Keywords", "attribute_value_mlt": [{"interim": "Machine Translation, Hierarchical Phrase-based Machine Translation (HPBSMT), Myanmar Sign Language (MSL), Myanmar SignWriting (MSW), Operation Sequence Model (OSM), Phrasebased Machine Translation (PBSMT)"}]}, "item_1583103120197": {"attribute_name": "Files", "attribute_type": "file", "attribute_value_mlt": [{"accessrole": "open_access", "date": [{"dateType": "Available", "dateValue": "2021-01-20"}], "displaytype": "preview", "download_preview_message": "", "file_order": 0, "filename": "Statistical Machine Translation between Myanmar Sign Language and Myanmar SignWriting.pdf", "filesize": [{"value": "3 MB"}], "format": "application/pdf", "future_date_message": "", "is_thumbnail": false, "licensetype": "license_3", "mimetype": "application/pdf", "size": 3000000.0, "url": {"url": "https://meral.edu.mm/record/7672/files/Statistical Machine Translation between Myanmar Sign Language and Myanmar SignWriting.pdf"}, "version_id": "fcab693a-cde4-4d91-8e88-47eb394fd5cd"}]}, "item_1583103131163": {"attribute_name": "Journal articles", "attribute_value_mlt": [{"subitem_pages": "Pages 65-72"}]}, "item_1583105942107": {"attribute_name": "Authors", "attribute_value_mlt": [{"subitem_authors": [{"subitem_authors_fullname": "Hnin Wai Wai Hlaing"}, {"subitem_authors_fullname": "Ye Kyaw Thu"}, {"subitem_authors_fullname": "Swe Zin Moe"}, {"subitem_authors_fullname": "Hlaing Myat Nwe"}, {"subitem_authors_fullname": "Ni Htwe Aung"}, {"subitem_authors_fullname": "Nandar Win Min"}, {"subitem_authors_fullname": "Hnin Aye Thant"}]}]}, "item_1583108359239": {"attribute_name": "Upload type", "attribute_value_mlt": [{"interim": "Publication"}]}, "item_1583108428133": {"attribute_name": "Publication type", "attribute_value_mlt": [{"interim": "Journal article"}]}, "item_1583159729339": {"attribute_name": "Publication date", "attribute_value": "2018-03-06"}, "item_title": "Statistical Machine Translation between Myanmar Sign Language and Myanmar SignWriting", "item_type_id": "21", "owner": "73", "path": ["1582963674932", "1597396989070"], "permalink_uri": "http://hdl.handle.net/20.500.12678/0000007672", "pubdate": {"attribute_name": "Deposited date", "attribute_value": "2018-03-06"}, "publish_date": "2018-03-06", "publish_status": "0", "recid": "7672", "relation": {}, "relation_version_is_last": true, "title": ["Statistical Machine Translation between Myanmar Sign Language and Myanmar SignWriting"], "weko_shared_id": -1}
Statistical Machine Translation between Myanmar Sign Language and Myanmar SignWriting
http://hdl.handle.net/20.500.12678/0000007672
http://hdl.handle.net/20.500.12678/0000007672e2babcc6-8106-45fe-9c08-26f5c3494699
cef27ac2-0593-41f7-ba09-29c0d721aeda
Name / File | License | Actions |
---|---|---|
![]() |
Publication type | ||||||
---|---|---|---|---|---|---|
Journal article | ||||||
Upload type | ||||||
Publication | ||||||
Title | ||||||
Title | Statistical Machine Translation between Myanmar Sign Language and Myanmar SignWriting | |||||
Language | en | |||||
Publication date | 2018-03-06 | |||||
Authors | ||||||
Hnin Wai Wai Hlaing | ||||||
Ye Kyaw Thu | ||||||
Swe Zin Moe | ||||||
Hlaing Myat Nwe | ||||||
Ni Htwe Aung | ||||||
Nandar Win Min | ||||||
Hnin Aye Thant | ||||||
Description | ||||||
This paper contributes the first evaluation of automatic machine translation between Myanmar Sign Language (MSL) and Myanmar SignWriting (MSW). The main motivation is to introduce SignWriting to the Myanmar Deaf society with the help of statistical machine translation. In this paper, we use our MSL-MSW corpus for general domain that contains a textual representation of MSL and its equivalent Myanmar SignWriting. The methods studied in this work were phrase-based, hierarchical phrase-based and the operation sequence model. In addition, two different segmentation schemes were studies, these were syllable segmentation and word segmentation for MSL. The performance of the machine translation systems was automatically measured in terms of BLEU and RIBES for all experiments. Our main findings were that operation sequence model gave the highest scores (37.54 BLEU and 0.8280 RIBES) for MSL to MSW translation and hierarchical phrase based machine translation gave the highest scores (52.79 BLEU and 0.8756 RIBES) for MSW to MSL translation. Generally, translation with word segmented MSL achieved better performance than syllable segmentation of MSL. Our 10-fold cross validation results produced promising results even with the limited training data and we expect this can be developed into a useful machine translation system as more data becomes available in the future. |
||||||
Keywords | ||||||
Machine Translation, Hierarchical Phrase-based Machine Translation (HPBSMT), Myanmar Sign Language (MSL), Myanmar SignWriting (MSW), Operation Sequence Model (OSM), Phrasebased Machine Translation (PBSMT) | ||||||
Journal articles | ||||||
Pages 65-72 |