རྫོང་ཁའི་སྐད་སྒྱུར་རིམ་ལུགས། Dzongkha Machine Translation

  Рет қаралды 13,618

རྫོང་ཁ་གོང་འཕེལ་ལྷན་ཚོགས། Dzongkha Development Commission

རྫོང་ཁ་གོང་འཕེལ་ལྷན་ཚོགས། Dzongkha Development Commission

Жыл бұрын

The Dzongkha Development Commission (DDC) has been pursuing Machine Translation works since a decade ago. It is not only Machine Translation tool but also other language technology tool such as the Automatic Speech Recognition (ASR) and and Text to Speech system (TTS). The biggest issue in developing such language technology is the lack of data or corpus in Dzongkha. Both monolingual and bilingual data are important. The DDC has been creating these data over a decade long and will pursue to create more. It is not only about creating Dzongkha data but a data with the latest standardized spellings and grammar and it must be well edited.
Several agencies are exploring to create tools, some are AI based, for Dzongkha but they get stuck when they realize that it cannot be done without data. Any agency in Bhutan who have created tools for Dzongkha were supported by DDC and without which they could not have achieved anything at all. For instance, the Department of Information Technology and Telecom (DITT / now GovTech Agency) used the Dzongkha fonts, locale, and keyboard layout of DDC to create Dzongkha support in Linux; the DHI and Omdena created a translation tool with corpus from DDC; the two bilingual dictionary Apps created by Diwas Puri were based on DDC dictionary data; and the Natural Language Processing (NLP) tools such as Machine Translation, ASR, and TTS created by College of Science and Technology (CST, Phuentsholing) was done in collaboration with DDC and funding support given by GovTech Agency.
The NLP tools created by CST and DDC jointly is a project aimed at creating AI based tools for Dzongkha to sustain and promote use of Dzongkha through ICT.
The current tools are available from nlp.cst.edu.bt/ and 'Google translate' has also provided translation service in Dzongkha.
In order to increase translation accuracy, Dzongkha data in several millions needs to be created and added to the data inventory. This is a huge task ahead but in the long run it will help all the Dzongkha learners and users.

Пікірлер
རྩ་བ་གསུམ། - Three gems of the country
2:24
རྫོང་ཁ་གོང་འཕེལ་ལྷན་ཚོགས། Dzongkha Development Commission
Рет қаралды 12 М.
སྤྱ་དང་བྱམོ། The monkey and the hen
25:40
རྫོང་ཁ་གོང་འཕེལ་ལྷན་ཚོགས། Dzongkha Development Commission
Рет қаралды 663 М.
🍉😋 #shorts
00:24
Денис Кукояка
Рет қаралды 1,8 МЛН
SHAPALAQ 6 серия / 3 часть #aminkavitaminka #aminak #aminokka #расулшоу
00:59
Аминка Витаминка
Рет қаралды 669 М.
བསམ་འཕེལ་ ༤ པ། Samphel 4
16:06
རྫོང་ཁ་གོང་འཕེལ་ལྷན་ཚོགས། Dzongkha Development Commission
Рет қаралды 58 М.
Sangwai Thabshey
15:16
MT AUDIO VISUAL
Рет қаралды 41 М.
བསམ་འཕེལ་༡ པ། Samphel-1
24:20
རྫོང་ཁ་གོང་འཕེལ་ལྷན་ཚོགས། Dzongkha Development Commission
Рет қаралды 592 М.
Let's Get Real with Kencho Tshering (Baby Drukpala)
48:10
Pema C
Рет қаралды 22 М.
Learn to speak Dzongkha (Bhutanese) with a native speaker!!
28:40
བསམ་འཕེལ་༣ པ། Samphel-3
20:39
རྫོང་ཁ་གོང་འཕེལ་ལྷན་ཚོགས། Dzongkha Development Commission
Рет қаралды 710 М.
India’s prostitution villages | DW Documentary
12:26
DW Documentary
Рет қаралды 10 МЛН
🍉😋 #shorts
00:24
Денис Кукояка
Рет қаралды 1,8 МЛН