CV
Education
- Preparatory, XJU, 09/2007-07/2008
- B.S, XJU, 09/2008-07/2012
- M.S, XJU, 09/2012-06/2015
- Ph.D, THU, 09/2015-06/2021
Work Experience
- Research Assistant, Laboratory of Multilingual Information Technology, Xinjiang Technical Institute of Physics and Chemistry, Chinese Academy of Sciences, 04/2023-Present
- Base: Urumqi, Xinjiang, China
- Project: New Frontiers of Natural Language Generation in Low-Resource Scenario
- NLP Senior Algorithm Engineer, Language Technology Lab, ALIBABA DAMO ACADEMY, 07/2021-03/2023
- Base: Hangzhou, Zhejiang, China
- Project: Multilingual Processing for International E-commerce Platforms DARAZ, LAZADA and AliExpress
- Contribution: owner of the DARAZ algorithm, both the FAQ and the Agent Intent Classification models. I contributed to the multilingual intelligent customer service system for AliExpress. I also proposed an improved code-mixing strategy and achieved better results in some languages of LAZADA with an average score of 4 points. I also presented some approaches for the pre-training part and achieved remarkably better results for the whole DARAZ system with 13 points (offline tests) and 9 points (online tests).
- Research Intern, Language Technology Lab, ALIBABA DAMO ACADEMY, 07/2020-10/2020
- Base: Hangzhou, Zhejiang, China
- Project: Urdu-Roman Transliteration for Smart Customer Service System of International E-commerce Platform
- Contribution: Contribution: I presented a hybrid model that can solve the main challenges (many-to-one mapping and inconsistent mapping) in this task. Moreover, the proposed approach is embedded in the dialog system of the international e-commerce platform DARAZ.
- CTO, Information Services Department, Beijing Silk Road Heli Investment Management Group Co.,Ltd, 07/2019-12/2019
- Base: Beijing, China
- Project: Multilingual E-commerce Platform
- Contribution: I had instructed our team to develop a multilingual e-commerce platform with front-end and back-end written by us instead of using other frameworks. The system is very flexible and extensible and can be effortlessly converted into an online teaching system or an online training system.
- Technical Advisor, MT Team, NewTranx Technology, 12/2018-05/2019
- Base: Shenzhen, Guangdong, China
- Project: NMT for Morphologically Rich Languages (Arabic-Chinese)
- Contribution: I had given some technical advice on training the NMT model for morphologically rich languages (Arabic, Turkish, and Urdu) between Chinese. We used a well-known NMT architecture (Transformer) with a commonly used neural network toolkit (Tensorflow). I also provided some processing tools for Arabic, Turkish and Urdu.
- Technical Advisor, MT Team, SinoVoice, 11/2017-11/2018
- Base: Beijing, China
- Project: NMT for Low-Resource Languages (Uyghur-Chinese)
- Contribution: I had advised the team of MT to implement the bidirectional Uyghur-Chinese NMT system efficiently and quickly. I had also provided both pre-processing and post-processing codes for Uyghur.
- Social Practice Intern, Information Team, Xinjiang Tourism Investment Management Co. Ltd, 06/2016-08/2016
- Base: Urumqi, Xinjiang, China
- Project: THU College Summer Social Practice Activities
- Contribution: In this project, I had submitted a technical report for the provision of the multilingual service. I also designed the main architectures for spoken translation by combining ASR, TTS and NMT for the Smart Tourism project.
- NLP Researcher (Apprenticeship), THUNLP Lab, Department of Computer Science and Technology, 09/2015 - 06/2021
- Base: Beijing, China
- Mentor: Prof. MaoSong Sun, Prof. Yang Liu
- Project: Cross-lingual Information Retrieval system and Low-Resource Languages NMT system
- Contribution: I have developed a cross-lingual information retrieval system between minority languages (Uyghur, Tibetan, Mongolian) and Chinese. I have also developed an NMT online demo for LRLs (demo).
- Visited Scholar, Machine Intelligence and Translation Lab, HIT, 09/2014-10/2014
- Base: Harbin, Heilongjiang, China
- Mentor: Dr. Conghui Zhu
- Project: Uyghur-Chinese Statistical Machine Translation (SMT)
- Task: Learn related techniques about SMT, such as learning open source systems MOSES, learn how to train, how to tune the model, how to optimize, and how to deploy the trained model on the web server. In the meantime, I learned the CRF model from Dr. Conghui Zhu. Besides, I also improved my Linux knowledge in their lab.
- Software Engineer, Information Team, Xinjiang Tiznap Information Technology Co. Ltd, 04/2013-11/2013
- Base: Urumqi, Xinjiang, China
- Project: Uyghur Search Engine
- Contribution: I had participated in the backend development of the Uyghur Information Retrieval System and gained some experience about the IR system.
- NLP Researcher (Apprenticeship), Xinjiang Multi-Lingual Information Technology Key Laboratory, 09/2012-06/2015
- Base: Urumqi, Xinjiang, China
- Mentor: Prof. Wushouer Silamu
- Project: Uyghur ASR, TTS, Spoken Translation systems
- Contribution: I had completed the Uyghur multi-script TTS model in collaboration with my teammates. In the meantime, we also proposed the Uyghur speech-driven ASR model. In the last year of my master’s degree, we had developed a bidirectional system for Uyghur-Chinese spoken translation by combining our individual models trained separately (Uyghur ASR, TTS and MT).
- Social Practice Intern, Information Team, Government office, 07/2010-08/2010
- Base: Kashgar, Xinjiang, China
- Project: XJU College Summer Social Practice Activities
- Contribution: In this project, we taught local farmers commonly used techniques and current preferential policies for farmers.
- Software Engineer (Apprenticeship), Xinjiang Multi-Lingual Information Technology Key Laboratory, 11/2009-06/2012
- Base: Urumqi, Xinjiang, China
- Mentor: Prof. Wushouer Silamu
- Project: Operation System (Localization), Software Development
- Contribution: I had taken part in the Win7 Operation System localization task. Besides, I had developed desktop applications (Education Systems) and Network applications (Government and Company multilingual Websites). Meanwhile, in the last year of my undergraduate, I designed Android APPs for learning Chinese.
Honors
- Outstanding student cadre of XJU, 2007-2008
- National Encouragement Scholarship, 2008-2009
- Excellent league member of XJU, 2008-2009
- National Encouragement Scholarship, 2009-2010
- Advanced Individual in Social Practice Intern of XJU, 10/2010
- Second prize in the program design competition of XJU, 06/2011
- National Scholarship (Undergraduate), 2010-2011
- Outstanding Student, 2010-2011
- Outstanding Graduates (Undergraduate), 06/2012
- Second-class scholarship for postgraduate, 09/2012, 09/2013, 09/2014
- National Scholarship (Postgraduate), 12/2014
- Third-class Social Practice Scholarship for Postgraduates of THU, 12/2016
- Second Prize of Science and Technology of Chinese Institute of Electronics, 01/2022
Students
- Research Intern
- Hongbin Na, Master, 06/2023-08/2023, The University of New South Wales (UNSW), Sydney, Australia
- Co-advised
- Dawulie Jinensibieke, Master, 08/2023-12/2023, University of Chinese Academy of Sciences, Beijing, China
- Wentao Xiao, Master, 06/2023-06/2024, Xinjiang University, Urumqi, Xinjiang, China
