The Chinese University of Hong Kong (Shenzhen) Li Haizhou/Professor Wang Benyou recruits doctoral students (2023Fall/2024Spring&Fall), research assistants and postdoctoral fellows in the direction of large models

team introduction

PhD Tutor

Li Haizhou

Li Haizhou , Executive Dean of the School of Data Science, Chinese University of Hong Kong (Shenzhen), Chair Professor of the President, Academician of the Singapore Academy of Engineering, Changjiang Scholar of the Ministry of Education, he is also a visiting professor at the National University of Singapore and an outstanding chair professor at the University of Bremen, Germany. He was a professor at Nanyang Technological University and National University of Singapore from 2006 to 2016, a visiting professor at the University of Eastern Finland in 2009, a visiting professor at the University of New South Wales in Australia from 2011 to 2016, and a visiting professor at the University of New South Wales from 2003 to 2016. In 2009, he served as the Chief Scientist and Research Director of the Institute of Communication and Information of the Singapore Agency for Science, Technology and Research. Professor Li is IEEE Fellow, ISCA Fellow, and AAIA Fellow. He was the editor-in-chief of top journals IEEE/ACM Transactions on Audio and Speech and Language Processing (2015-2018). He has also been the chairman of several large international academic conferences, including ACL 2012, INTERSPEECH 2014, and ICASSP 2022.

Homepage https://colips.org/~eleliha/

Wu ruling

Wu Zhizheng is an associate professor at the Chinese University of Hong Kong (Shenzhen). Prior to this, he received a Ph.D. from Nanyang Technological University in 2015, and engaged in academic research and technical leadership at Meta (formerly Facebook), JD.com, Apple, University of Edinburgh, Microsoft Research Asia and other institutions. He has won the Best Student Paper Award at INTERSPEECH 2016 and the Best Paper Award at the 2012 Asia Pacific Signal and Information Processing Association Annual Summit. Professor Wu led the development of the speech synthesis open source system Merlin, initiated and organized the first international evaluation of voiceprint recognition and deception detection (ASVspoof), the first international evaluation of speech conversion, and organized the 2019 international evaluation of speech synthesis (Blizzard Challenge 2019) .

Homepagehttps ://drwuz.com

Wang Benyou

Benyou Wang is an assistant professor at The Chinese University of Hong Kong (Shenzhen). His main research directions are natural language processing (NLP), applied machine learning, and information retrieval. So far, he has won the Best Paper Nomination Award at SIGIR 2017, the Best Interpretable NLP Paper at NAACL 2019, the Best Paper at NLPCC 2022, the Huawei Spark Award, and the EU Marie Curie Scholarship. He led the development of the Phoenix multilingual model and the Huatuo GPT model. The Phoenix model ranked fifth on the SuperCLUE Nirvana Ranking ¹ manual evaluation . The doctor evaluated that Huatuo GPT has surpassed ChatGPT 3.5, and is currently accelerating its iteration.

Homepage https://wabyking.github.io/old.html

School and team advantages

The team has close ties with the industry, and has long-term cooperation with MSRA/Huawei/Amazon/Microsoft Montreal/MILA/Zhiyuan/Tencent/Ali, providing abundant internship and visiting opportunities, and encouraging members to go to world-class universities and research institutions for exchanges (Singapore National University/Hong Kong University/University of Montreal/MILA/University of Copenhagen, etc.).

The school implements the same teaching system as in North America, with a five-year schooling period (four years for a master’s degree), and a degree from the Chinese University of Hong Kong after graduation; an open international campus, good and convenient living and learning facilities; outstanding students can apply for generous subsidies;
Due to the expensive training of doctoral students (annual tuition and living expenses are close to 20W), doctoral students can focus on academic issues to the greatest extent and reduce some trivial interference (specially hired people to be responsible); because the training funds are too high and similar to the North American funding mechanism, doctoral students are more Participate less in horizontal projects.
First-class doctoral supervisors, cutting-edge doctoral topics, strong scientific research atmosphere, free cooperation and communication within the team without any scruples; welcome to maintain cooperation with domestic and foreign counterparts, and support domestic and foreign meetings and exchanges;
Cooperate closely with first-class well-known companies/research institutions in the industry, have abundant computing resources, try not to let computing resources limit imagination, and give full play to people's creativity; currently the team has sufficient A100 computing power, massive data and other research resources , and has a large number of external deployment resources, and hopes to display the research content in the form of an online demonstration website;
The advantage of the team is that there are energetic young PIs and senior PIs who can provide multi-faceted and multi-level guidance.
Professor Liu Qun (ACL Fellow) and Dr. Jiang Xin from Huawei Noah's Ark Laboratory, Professor Zong Chengqing from the Institute of Automation, Chinese Academy of Sciences (ACL/IEEE Fellow), Professor Wan Xiang from Shenzhen Institute of Big Data, and Professor Xu Jinchao from KAUST (SIAM Fellow) ) and other close cooperation, there is a joint training doctoral program.
Support graduates to go to academia/industry for development; earnestly hope to train independent doctoral students who can be independent after graduation and have the potential to become independent PIs in first-class universities; undertake top-level conferences in related fields at home and abroad, and support students to undertake conference services, Facilitate early career starts in academia and industry.

about research

research direction

Construction of large-scale language models: Systematic research mainly focuses on core links such as pre-training, instruction fine-tuning, and reinforcement learning. At the same time, it also focuses on plug-ins, retrieval enhancement generation, and other research directions to enhance the capabilities of large models;
Optimization and application of large-scale language models: including but not limited to modularization of large-scale models, knowledge injection, prompt engineering (Prompt Engineering), evaluation, etc., but also research on human-computer interaction based on large-scale language models, generated content detection, decision-making reasoning, etc.;
Research on multi-modal large models: realize the unified understanding and generation of language models for multiple data formats, including but not limited to pictures, audio, tables, codes, and videos;
The civilianization of large models: making large-scale language models more available, including improving training and deployment efficiency, lowering the threshold for data collection, etc. (see our Phoenix large model);
Industrialization of large models: Integrate big data models into all aspects of life to improve work efficiency and quality of life, including applications in the medical field (see our Hua Tuo GPT).

Existing research examples

We have developed several large models ( Phoenix Large Model , Huatuo GPT ), welcome to experience, or you can experience it by searching the "Shenxian Lake" applet on WeChat.

research philosophy

The team's scientific research advocates "discovering new phenomena, defining new problems, and designing new paradigms". The research philosophy is as follows:

The research team adheres to the scientific research concept of "Less is more" and does simple and effective work;
Do impactful research, papers as a way (rather than an end) to showcase our work, do interesting or useful papers, and strive to contribute to the community;
Doing research on entropy reduction will make practitioners more enlightened, not more confused;
Do general and in-depth research;
Do engineering, product and scientific research together, and use scientific research methods to solve problems that engineering and products cannot solve.

Recruitment requirements

Professors Li Haizhou and Wang Benyou of The Chinese University of Hong Kong (Shenzhen) plan to recruit 4 doctoral students, 3 research assistants and 6 postdoctoral fellows. The research direction is as introduced above. Our team also recruits unlimited number of master students, visiting students/scholars. Students with backgrounds in computer science, natural language processing, speech processing, data science, machine learning, information retrieval, etc. are welcome to apply.

Doctoral Recruitment

basic requirements

Enrollment target: direct Ph.D. (undergraduate students), postgraduate students who intend to study for a Ph.D. (application system is adopted, no postgraduate entrance examination is required)

Bachelor degree or above (the application system does not require postgraduate entrance examination), please refer to the school background of students admitted in previous years ( https://sds.cuhk.edu.cn/student-search ), those with computer science and other related professional backgrounds are preferred;
Excellent grades during school, solid foundation in mathematics, good English reading, writing and programming skills;
Those who have research experience in large-scale language models, dialogue systems, text generation, and multimodality, and have corresponding papers in well-known conferences or journals at home and abroad are preferred;
Love scientific research, have a sense of responsibility, good communication skills, professionalism and teamwork spirit;
Obtained a degree or certificate of study in Hong Kong or an English-speaking country; TOEFL (no less than 550 points in the written test, no less than 213 points in the computer-based test; no less than 79 points in the internet-based test); IELTS (no less than 6.5 points in the academic category).

Application Procedures

The doctoral application process is as follows:

Please prepare a complete Chinese/English resume (please attach grades and paper publication records) and send it to Dr. Jiang Feng, a member of the team, at the email address: [email protected], with the title of the email indicating: PhD application + personal name + interested research direction;
Interview to understand the candidate's background, research ability and research interests;
College admissions process.

Important time:

Enrollment time: September 2023 (fall), January 2024 (spring), September 2024 (fall)
Admission Deadline: There are still a small number of places for autumn admission, which are open to particularly outstanding students. There is no deadline for the deadlines for Spring 2024 and Fall 2024, the earlier the better.

For details of doctoral application in Chinese University of Hong Kong (Shenzhen), please refer to:

Ph.D. in Data Science: https://sds.cuhk.edu.cn/en/phd-programmes\
Ph.D. in Computer and Information Engineering: http://sse-mphil-phd.cuhk.edu.cn/en/basic/249
Ph.D. in Computer Science: https://sds.cuhk.edu.cn/en/phd-programmes-CSE

Full-time research assistant (Research Assistant) and postdoc

basic requirements

research assistant

Bachelor’s degree or above (Master’s and Ph.D. students can apply);
Have certain English reading, writing and programming skills;
Passionate about scientific research, full of responsibility and teamwork spirit.

postdoc

A PhD defense is required;
Candidates with computer science, mathematics and other related professional background are preferred;
Encourage freedom to explore and do world-class work.

Job Responsibilities, Salary and Benefits

Job responsibilities, salary and benefits are as follows:

Engage in scientific research and write scientific research papers;
Competitive salary, determined according to personal qualifications and experience; postdoctoral fellows can apply for the Rising Star postdoctoral program, the highest annual salary can reach 74W RMB, plus 30W research start-up funds, see https://sds.cuhk.edu.cn/ for details article/952 ;
Excellent research assistants can apply for a master's or doctoral degree in the School of Data Science, The Chinese University of Hong Kong, Shenzhen;
Postdoctoral fellows can assist in guiding undergraduates, masters and doctoral students in the team, and write strong recommendation letters to support the next stop of academic career;
Returned overseas students who meet the conditions can apply for the talent plan;
Most of the postdoctoral fellows from the School of Data Science joined the domestic 985 and top 211 after leaving the station, and many of them obtained tenured teaching positions.

Application Procedures

The application procedure is as follows:

Please prepare a complete resume in Chinese/English (please attach grades and publication records) and send it to Dr. Feng Jiang, a member of the team, by email: [email protected]. The title of the email indicates: the position you are applying for + your name;
Suitable candidates will be interviewed by team members to understand the candidate's background, ability and research interests;
This vacancy is valid until a suitable candidate is recruited.

background introduction

The Chinese University of Hong Kong (Shenzhen) is a university approved by the Ministry of Education to inherit the educational philosophy and academic system of the Chinese University of Hong Kong. At present, more than 8,000 outstanding students from all over the world are studying in CUHK-Shenzhen. After nine years of development, the discipline construction of the university has been gradually improved, and more than 400 internationally renowned outstanding scholars and researchers have been recruited from around the world, including 5 Nobel Prize winners, 2 Turing Award winners, and Fields Medal winners. 1, nearly 30 academicians from various countries (including 10 full-time), nearly 60 national specially-appointed experts, and nearly 40 fellows of ACM/IEEE and other associations. At present, 100% of the teachers introduced have teaching or research experience in world-class universities. The university has been the university with the highest admission score among institutions in Guangdong Province for six consecutive years, and graduates are awarded a degree certificate from the Chinese University of Hong Kong.

The Chinese University of Hong Kong (Shenzhen) School of Data Science currently has 62 faculty members (including 15 part-time employees), including 4 principal attendance professors, 10 principal lecture professors, 12 professors, 11 associate professors, 23 assistant professors, 1 lecturer. The faculty includes 1 foreign double academician of the Royal Canadian Academy of Sciences and the Chinese Academy of Engineering, 1 double academician of the Royal Canadian Academy of Sciences and the Canadian Academy of Engineering, 1 academician of the Singapore Academy of Engineering, 1 fellow of INFORMS, the International Electrical and Electronic Engineering There are 4 IEEE fellows, 1 IASCYS fellow of the International Academy of Systems and Control Sciences, 2 IMS fellows of the International Society for Mathematical Statistics, 2 SIAM fellows of the American Society for Industrial and Applied Mathematics, 1 ISCA fellow of the International Speech Communication Society, There are 9 national high-level talents and 7 provincial and municipal high-level talents. At present, the PhD students admitted by the School of Data Science are all from well-known universities, such as Tsinghua University, Peking University, Shanghai Jiaotong University, Zhejiang University, Nanjing University, Renmin University of China, University of Science and Technology of China, Wuhan University, Tongji University, Chinese University of Hong Kong ( Shenzhen), Southern University of Science and Technology, etc. Undergraduate majors include mathematics and applied mathematics, computer science, information science, industrial engineering, etc. Among them, 42% of students' undergraduate grades rank in the top 5% of their majors, and 80% of students' undergraduate grades rank in the top 10% of their majors.

https://www.superclueai.com/ ↩︎