一、項目名稱

A New Paradigm for Efficient Statistical Inference via Deep Representation Learning and Double Sampling

二、項目簡介

In the era of big data, obtaining high-quality "gold standard" data (e.g., precise clinical diagnoses) is often costly and procedurally complex, resulting in very limited sample sizes for such validation data in practice. Meanwhile, large-scale and easily accessible surrogate data (e.g., routine monitoring indicators), despite containing measurement errors or incomplete information, are rich in auxiliary information. How to efficiently and robustly leverage both types of data for statistical inference poses a significant challenge in contemporary statistics and data science.This project innovatively integrates the powerful representation learning capability of deep neural networks (DNNs) with the statistical efficiency theory of double sampling. We construct a zero-mean correction term derived from the surrogate data and incorporate it into the parameter estimation based on the validation sample. This approach achieves variance reduction, thereby improving estimation efficiency, and enables robust statistical inference—meaning that the inference results are robust to the surrogate data model. This method can be widely applied in medicine (e.g., disease diagnosis using small-sample accurate models combined with large-sample surrogate data) as well as in other complex data scenarios such as image analysis, offering a new pathway for efficient modeling under data scarcity.

在大數據時代，獲取高質量“金標準”數據（如精確臨床診斷）往往成本高昂且過程複雜，導致實際應用中這類驗證樣本的規模十分有限。與此同時，大規模、易獲取的替代數據（如常規監測指標）雖然包含測量誤差或信息不完整
，卻蘊含著豐富的輔助信息。如何充分利用這兩類數據實現高效
、穩健的統計推斷
，是當前統計學與數據科學麵臨的重要挑戰。

本項目創新性地將深度神經網絡（DNN）強(qiang)大(da)的(de)表(biao)示(shi)學(xue)習(xi)能(neng)力(li)與(yu)雙(shuang)重(zhong)抽(chou)樣(yang)的(de)統(tong)計(ji)效(xiao)率(lv)理(li)論(lun)相(xiang)結(jie)合(he)，構(gou)建(jian)一(yi)個(ge)源(yuan)於(yu)替(ti)代(dai)數(shu)據(ju)的(de)零(ling)均(jun)值(zhi)校(xiao)正(zheng)項(xiang)
，並(bing)將(jiang)其(qi)整(zheng)合(he)到(dao)基(ji)於(yu)驗(yan)證(zheng)樣(yang)本(ben)的(de)參(can)數(shu)估(gu)計(ji)中(zhong)，實(shi)現(xian)方(fang)差(cha)縮(suo)減(jian)
，從(cong)而(er)提(ti)升(sheng)估(gu)計(ji)效(xiao)率(lv)，並(bing)實(shi)現(xian)穩(wen)健(jian)的(de)統(tong)計(ji)推(tui)斷(duan)——即推斷結果對替代數據模型具有魯棒性。該方法可廣泛應用於醫學領域（如基於小樣本精確模型與大樣本替代數據的疾病診斷）
，yijituxiangfenxidengqitafuzashujuchangjing，weishujuxiquetiaojianxiadegaoxiaojianmotigongxinsilu。chushangshujutideketiwai

，lianheketizujianggenjushenqingrendejiaoyubeijingheyiyoudeyanjiujichuzhidaoqizaiyinguotuiduan、統計學習
、AI for Statistics or Statistics for AI、AI for Finance上探索新穎的前沿課題。

三、PIs at SUSTech & CUHK

Prof. Xuejun JIANG （蔣學軍）
，南方科技大學統計與數據科學係

Faculty Profile: Professor Xuejun Jiang is a tenured Associate Professor, Deputy Department Head, and doctoral supervisor at the Department of Statistics and Data Science, Southern University of Science and Technology (SUSTech). He earned his Ph.D. in Statistics from The Chinese University of Hong Kong (CUHK) in 2009, followed by postdoctoral research there (2009–2010), and joined SUSTech in 2013. He was honored as a recipient of the Shenzhen Overseas High-Level Talent Peacock Program (2016) and a Shenzhen Outstanding Teacher (2018). He has led over 10 research projects funded by the NSFC, Guangdong Provincial Natural Science Foundation, and Shenzhen Basic Research Program, etc. Xuejun Jiang’s research interests include complex data analysis, feature extraction, statistical inference, financial and applied statistics, and machine learning (e.g., transfer learning, representation learning, auxiliary learning, conformal inference, etc.). He has published over 60 SCI/SSCI papers in leading journals including Biometrika, Bernoulli, Statistica Sinica, JBES, The Econometrics Journal, Science China-Mathematics, and Scientia Sinica-Mathematica with two authorized patents and one English textbook.

Professor Jiang also serves as Vice Chairperson of the Educational Statistics and Management Branch and Secretary-General of the Multivariate Analysis Application Committee under the Chinese Association for Applied Statistics (CAAS). 教授簡介: 蔣學軍

，南方科技大學統計與數據科學係研究員(長聘副教授), 係負責人，博士生導師。他於2009年博士畢業於香港中文大學統計係

，2009-2010在港中文從事博士後研究， 2013年07月加入南方科技大學
，入選深圳市海外高層次人才孔雀計劃（2016），深圳市優秀教師（2018），主持和完成國家（廣東省）自然科學基金
、深圳市基礎研究麵上項目等10餘項。其研究方向和興趣涉及複雜數據分析、特征提取、統計推斷
，金融/應用統計
，機器學習(遷移學習及表征學習、輔助學習、共形推斷與預測)等，已在統計學頂級期刊Biometrika及Bernoulli, Statistica Sinica,JBES,The Econometrics Journal, Science China-Mathematics,中國科學數學等統計學及計量經濟學國內外一流學術期刊上發表SCI&SSCI論文60餘篇，授權專利2項及出版英文教材一部。國內學會任職主要有中國現場統計研究會-教育統計與管理分會副理事長，多元分析應用專業委員會秘書長等。Prof. Xinyuan Song (宋心遠)
，香港中文大學統計與數據科學係

Research Group Website: http://www.sta.cuhk.edu.hk/xysong/

Faculty Profile: Song Xinyuan is a Professor in the Department of Statistics and Data Science at The Chinese University of Hong Kong (CUHK), a Fellow of the Institute of Mathematical Statistics (IMS Fellow), as well as an Elected Member of the International Statistical Institute (ISI). Her research interests span a broad range of areas, including include latent variable models, Bayesian methods, survival analysis, nonparametric and semiparametric methods, causal inference, and statistical computing, etc. She has published over 230 papers in leading international journals in statistics and related fields. Additionally, Professor Song serves as an Associate Editor for several top-tier international journals in statistics and psychometrics, such as JASA (Journal of the American Statistical Association), Biometrics, and Psychometrika, etc.

教師簡介：宋心遠
，香港中文大學統計與數據科學係教授，國際數理統計學會會士（IMS Fellow），國際統計協會當選會員(ISI Elected Member)。她的研究興趣廣泛
，包括潛變量模型、貝葉斯方法、生存分析
、非參數與半參數方法、因果推斷及統計計算等。目前已在統計學及相關學科國際一流期刊上發表論文230餘篇。宋心遠教授現任多個國際統計與計量心理學期刊的副主編，包括JASA

，Biometrics，Psychometrika等。

四
、崗位要求

We are hiring 1 postdoctoral fellow. The details are listed below.

課題組現公開招聘博士後1名，具體崗位信息如下：

01崗位要求

1. Hold a PhD degree (or complete a PhD program in 2026) in Statistics, Mathematics, Data Science, Computer Science or other related areas. Graduates from renowned overseas universities or "985" universities in China are preferred. 2. Proficiency in R,Python/Matlab or other computer languages.3. Good knowledge and strong research abilities in statistical/mathematical methodology, theory and implementation, preferable on high-dimensional data analysis, complex modeling, or image processing, as well as those who have a foundational understanding and interest in AI for Statistics, Statistics for AI, or AI+Statistics for Public Health or Finance. 4. Excellent English writing skills are required. Prior experience in writing research papers or grant proposals is preferred. 5. Good communication and presentation skills in both English and Chinese. 6.This project is a collaboration between the research group projects of Southern University of Science and Technology and The Chinese University of Hong Kong (CUHK). During the postdoctoral period, candidates may have the opportunity to conduct short-term visits and exchanges at CUHK.The postdoctoral position must comply with the postdoctoral position management regulations of Southern University of Science and Technology. Specific cooperation details are to be discussed in person

1.獲得或即將獲得統計學、數學、數據科學、計算機或其他相關學科的博士學位（博後的要求），境外名校或“985”高校相關專業博士生優先
；

2.精通R,Python/Matlab或其他至少一種計算機語言；

3.有較強的統計學/數學方法和理論基礎知識和實踐能力
；有高維複雜數據分析、複雜模型或圖像處理研究經驗者優先或對AIforStatistics,StatisticsforAI以及AI+StatisticsforPublicHealthorFinance有基礎和興趣的優先；

4.具有較強英文寫作能力，有論文或項目書等寫作經驗者優先；

5.具有良好的溝通能力和展示能力；

6.本ben項xiang目mu為wei南nan方fang科ke技ji大da學xue與yu香xiang港gang中zhong文wen大da學xue課ke題ti組zu項xiang目mu之zhi間jian的de合he作zuo，博bo士shi後hou在zai站zhan期qi間jian可ke允yun許xu到dao香xiang港gang中zhong文wen大da學xue進jin行xing短duan期qi交jiao流liu訪fang問wen，但dan博bo士shi後hou崗gang位wei須xu遵zun循xun南nan方fang科ke技ji大da學xue博bo士shi後hou崗gang位wei管guan理li規gui定ding

，具ju體ti合he作zuo方fang式shi麵mian議yi。

02崗位職責

1. Undertake research related to the project.

2. Help to prepare research proposals.

3. Help on other research activities.

1.進行與本課題相關的科研工作；

2.協助課題組申報各類科研課題及承擔相應的科學研究任務；

3.協助完成課題組的其他日常工作。

03待遇與福利

1.The postdoctoral employment period is two years, with a comprehensive annual salary starting from 330,000 RMB (before tax, including living subsidies for postdocs in station from Guangdong Province and Shenzhen City). Exceptionally outstanding candidates can apply for the President's Distinguished Postdoctoral Fellowship, with an annual salary of up to 500,000 RMB or more (including provincial and municipal subsidies). 

2. Guangdong Province provides a total living subsidy of 300,000 RMB (before tax) per person for eligible postdocs in station. Shenzhen City provides a total living subsidy of 120,000 RMB (before tax) per person for eligible postdocs in station, with a funding period of 24 months. 

3. During the station period, postdocs can rely on the school to apply for Shenzhen public rental housing. Postdocs who do not use Shenzhen public rental housing through the school can enjoy a pre-tax housing subsidy of 2,800 RMB/month for two years. 

4. Possess an excellent working environment and opportunities for domestic and international cooperative exchange. Postdocs enjoy a total of 25,000 RMB in academic exchange funding during their two-year station period. 

5. The research group can assist eligible postdocs in applying for postdoctoral talent projects. Upon approval, a maximum total subsidy of 1 million RMB can be enjoyed (cannot be enjoyed simultaneously with provincial and municipal subsidies). 

6. For postdocs who stay in (or come to) Shenzhen for full-time work within 6 months after leaving the station and sign a labor (employment) contract of 3 years or more with enterprises or public institutions, the Shenzhen Municipal Government will provide a living subsidy of 360,000 RMB per person for coming to Shenzhen after leaving the station. 

7. Outstanding postdoctoral personnel who obtain the Postdoctoral Innovative Talent Support Program or the special funding from the China Postdoctoral Science Foundation during the station period, and sign a labor (employment) contract of 3 years or more with this city within 6 months after leaving the station, the Shenzhen Municipal Government will provide 1:1 matching funds according to national funding standards, up to a maximum of 300,000 RMB. 

8. For those who win the Gold, Silver, or Bronze awards in the National or Guangdong Postdoctoral Innovation and Entrepreneurship Competition, and sign a labor (employment) contract of 3 years or more with this city within 6 months after leaving the station, the Shenzhen Municipal Government will provide 1:1 matching innovation and entrepreneurship rewards according to the national and provincial reward amounts, up to a maximum of 200,000 RMB. 

9. According to Article 39 of the "Regulations on the Administration of Postdoctoral Work in Shenzhen", the funding items in these regulations and the living subsidy for newly introduced postdoctoral talents in Shenzhen (100,000 RMB) shall not be enjoyed repeatedly.

1.博士後聘用期兩年
，綜合年薪33萬元起（稅前，含廣東省及深圳市博士後在站生活補助），特別優秀候選人可以申請校長卓越博士後

，年薪可達50萬元以上（含省市補助）。

2.廣東省對符合條件的在站博士後發放每人總額30萬元（稅前）的生活補助，深圳市對符合條件的在站博士後發放每人總額12萬元（稅前）的生活補助
，資助期為24個月。

3.在站期間，可依托學校申請深圳市公租房
，未依托學校使用深圳市公租房的博士後，可享受兩年稅前2800元/月的住房補貼。

4.擁有優良的工作環境和境內外合作交流機會
，博士後在站期間享受兩年共計2.5萬學術交流經費資助。

5.課題組可協助符合條件的博士後申請博士後人才項目。獲批最高可享受總計100萬元補貼（與省市補助不同時享受）。

6.博士後出站後6個月內留（來）深全職工作且與企事業單位簽訂3年以上勞動（聘用）合同的
，深圳市政府給予每人36萬元出站來深生活補助。

7.在站期間獲得博士後創新人才支持計劃或中國博士後科學基金特別資助的優秀博士後人員，且出站後6個月內與本市簽訂3年以上勞動（聘用）合同的，深圳市政府再按照國家資助標準給予1：1經費資助


，最高不超過30萬元。

8.獲得全國或廣東省博士後創新創業大賽金獎
、銀獎
、銅獎的
，且出站後6個月內與本市簽訂3年以上勞動（聘用）合同的，深圳市政府再按照國家和省獎勵金額給予1：1創新創業獎勵，最高不超過20萬元。

9.根據《深圳市博士後工作管理規定》第三十九條規定，該規定資助項目與深圳市新引進博士人才生活補貼（10萬元）不重複享受。

04聯係方式（before31Aug,2026）

To apply for the position, please send the following information to Prof. Jiang（（點擊查看））and Prof. Song（（點擊查看）） with the title “SUSTECH & CUHK JOINT RESEARCH PROJECT -position-your name-your major”.1.Resume (with a complete list of publications and transcripts).2. The full manuscript of 2 representative publications. 3. Other research outputs such as books, patents, etc.有意向者請將個人詳細簡曆(包括成績單和已發表文章的完整列表)、代表性學術成果等整合為一個PDF文件
，郵件發送至蔣學軍老師（（點擊查看） ）和宋心遠老師（（點擊查看））郵件標題請注明：SUSTech & CUHK聯合研究項目-崗位-姓名-專業。

信息來源於網絡，如有變更請以原發布者為準。

來源鏈接：

https://mp.weixin.qq.com/s/0NAJKllmIjp3-02K7nj-4g

更多最新博士後招收資訊請關注：

【高才博士後】網站→https://boshihou.gaoxiaojob.com

【高才博士後】微信公眾號→微信號：bshjob