News

One full paper accepted to IJCAI2021 (accepted rate=13.9%)

One long paper accepted to EMNLP 2020 (oral).

One paper accepted to ICLR 2019.

个人简介

Wanyun Cui is an assistant professor in the school of information management and engeering at Shanghai University of Finance and Economics. He lead the AI&Finance research group here. He is also the director of Xiaocui Robot (Wechat: 小Cui问答) and KBQA(demo). He received his PhD's degree in the school of computer science at Fudan University in 2017. His PhD thesis <Research of Key Technologies for Question Answering over Knowledge Bases> won the "2017 ACM China Excellent Doctoral Thesis" nomination (top 4 in China), and "2017 ACM Shanghai Excellent Doctoral Thesis award" (top 2 in Shanghai). His research interests include question answering and knowledge graphs. He has been working on QA systems at Microsoft Research Asia, Baidu, and Xiaoi Robot since 2012. He has published related papers as the first author in VLDB, IJCAI, AAAI, and SIGMOD. He received his bachelor's degree from Fudan University in 2013. He is also the winner of Fudan academic star award and Shanghai outstanding graduate award.

崔万云是上海财经大学信息管理与工程学院的助理教授。他仅以4年时间就获得了复旦大学5年制博士学位，同时获得了复旦大学最高博士荣誉“复旦学术之星”（计算机方向唯一）。其博士论文《基于知识图谱的问答系统关键技术研究》被授予了“2017ACM中国优秀博士生论文提名奖”，和“2017ACM上海优秀博士生论文奖”(top 2 in 上海)。和他的研究兴趣包括自然语言问答和知识图谱。自2012年起，他分别在微软亚洲研究院、百度深度问答小组和小i机器人等公司从事问答系统、知识图谱相关研究。他已经在ICLR，PVLDB，SIGMOD，EMNLP，AAAI，IJCAI等顶级人工智能、数据库会议上发表多篇第一作者论文。他还曾获得复旦大学毕业生之星、上海市优秀毕业生等奖项。

毕业生去向

• 郑光煜  硕士  蚂蚁金服
• 王乐  硕士 猿辅导
• 蔡令予  硕士  阿里巴巴

I am recruiting highly self-motivated and hardworking students. If you enjoy doing challenging research and want to devote your career to research, please do not hesitate to contact me.

教授课程

科研项目

问答系统:

1. KBQA: 作为主要负责人 基于知识图谱和问答语料的自然语言问答系统，在QALD-3上正确率76%，超过所有同类系统。News 已经支持中文问答，参见公众号：小Cui问答。
2. 小度机器人: 负责实体类答案深度问答技术，百度深度问答项目，在一站到底数据集上准确率从81%提升到85%。
3. 小i机器人: 负责问答系统架构重构，小i机器人公司项目，将其问答技术核心从传统文本挖掘信息检索，向知识图谱和深度学习方向引导。

其它:

1. 中文同义词库: 作为主要负责人提供368万中文实体-别称的高质量实体同义词库。利用开放的互联网数据及深度学习模型。
2. 人人猎头钢铁侠&复旦生涯: 作为主要负责人对简历和职位做语义理解和智能匹配。从百万级别简历/职位数据中在1秒内实现高精确度的匹配。应用于人人猎头APP和复旦生涯微信号。
3. 两会舆情分析: 负责13年两会微博热点词汇分析,工作被新民晚报及解放日报报导。
4. AffineSWOptimization: 作为主要负责人 使用CPU级的SSE2技术加速Smith-Waterman算法，在Topcoder的marathon match中排名第八。

Projects
Question Answering Systems:

1. KBQA: Main Contributor Project in Microsoft Research Asia. A QA system over knowledge bases. Its precision on QALD-3 is 76%, which beats other state-of-the-art competitors. News Supporting Chinese QA now.
2. Deep QA: in charge of factoid question answering Project in Baidu Inc. (Chinese largest search engine company) The precision increases from 81% to 85% on the Chinese YiZhanDaoDi dataset.
3. Xiaoi Robot: in charge of qa system reconstruction Project in Xiaoi Robot Inc. Reconstructing its qa system framework from traditional information retrieval techniques to deep learning and knowledge graph techniques.

Other Projects:

1. Chinese mention2entity: Main Contributor Providing 3.68 million mention2entity facts from web knowledge extraction and deep learning.
2. Iron Man Project in Shanghai Zhongpin Inc. & Fudan Career Project in Fudan Univeersity: Main Contributor Rank and sort millions of resume descriptions and job descriptions with nlp techniques.
3. Public Sentiment Analysis for the Chinese National Sessions: in charge of hot keyword analysis The result has been published in two Chinese well-known newspapers: Xinmin Evening News, and Liberation Daily.
4. AffineSWOptimization: Main Contributor Using CPU leveled instructions, SSE2, to accelerate the Smith-Waterman algorithm. Rank 8th in the Topcoder marathon match.

教育和工作经历

实习经历 

1. 2014.7-2014.11 百度深度问答项目（小度机器人），实体类答案问答技术
2. 2012.1-2012.11：微软亚洲研究院,构建基于大型知识库的问答系统。

Experiences:

1. Jul 2014 - Nov 2014: Deep QA project, Baidu Inc. (Chinese largest search engine company)
2. Jan 2012 - Nov 2012: Microsoft Research Asia, working on the construction of QA system over large scale knowledgebase, with Dr. Haixun Wang

研究成果

1. Wanyun Cui, Guangyu Zheng, Wei Wang, Zero-shot domain adaptation for natural language inference by projecting superficial words out, Knowledge-Based Systems, (IF=5.92)
2. Wanyun Cui, Sen Yan, Isotonic Data Augmentation for Knowledge Distillation, IJCAI 2021, CCF Rank A Conference
3. Wanyun Cui, Guangyu Zheng, Wei Wang, Unsupervised Natural Language Inference via Decoupled Multimodal Contrastive Learning, EMNLP 2020 (oral), [code]
4. Wanyun Cui, Guangyu Zheng, Zhiqiang Shen, Sihang Jiang, Wei Wang, Transfer Learning for Sequences via Learning to Collocate, ICLR 2019, [code]
5. Wanyun Cui, Yanghua Xiao, Haixun Wang, Yangqiu Song, Seung-won Hwang, Wei Wang, KBQA: Learning Question Answering over QA Corpora and Knowledge Bases, (PVLDB 2017), CCF Rank A Conference
6. Bo Xu, Yong Xu, Jiaqing Liang, Chenhao Xie, Bin Liang, Wanyun Cui, Yanghua Xiao, CN-DBpedia: A Never-Ending Chinese Knowledge Extraction System, (IEA/AIE 2017 )
7. Wanyun Cui, Yanghua Xiao, Wei Wang, KBQA: An Online Template Based Question Answering System over Freebase, (IJCAI 2016), CCF Rank A Conference, demo
8. Wanyun Cui, Xiyou Zhou, Hangyu Lin, Yanghua Xiao, Haixun Wang, Seung-won Hwang, Wei Wang, Verb Pattern: A Probabilistic Semantic Representation on Verbs, (AAAI 2016), CCF Rank A Conference PDF interface
9. Wanyun Cui, Yanghua Xiao, Haixun Wang, Wei Wang, Local Search of Communities in Large Graphs, (SIGMOD 2014), CCF Rank A Conference PDFPPT
10. Wanyun Cui, Yanghua Xiao, Haixun Wang, Yiqi Lu, and Wei Wang. "Online Search of Overlapping Communities.", (SIGMOD 2013), CCF Rank A Conference PDF, PPT
11. Deqing Yang, Yanghua Xiao, Hanghang Tong, Wanyun Cui, Wei Wang, Towards Topic Following in Heterogeneous Information Networks, (ASONAM 2015)
12. Hui Wang, Wanyun Cui, Yanghua Xiao, Hanghang Tong, Robust Network Construction against Intentional Attack, (BigComp 2015), Invited Paper
13. Yaoliang Chen, Ji Hong, Wanyun Cui, Jacques Zaneveld, Wei Wang, Richard Gibbs, Yanghua Xiao and Rui Chen, CGAP-align: A High Performance DNA Short Read Alignment Tool, Plos One, 2013, 8(4): e61033, SCI, IF=4 PDF
14. Yanghua xiao, Ji Hong, Wanyun Cui, Zhenying He, Wei Wang, Guodong Feng, Branch Code: An Efficient Labeling Scheme for Query Answering on Trees, IEEE 28th International Conference on Data Engineering(ICDE 2012) CCF Rank A Conference.

荣誉奖励

所获奖项 

1. ACM中国优秀博士论文奖提名
2. ACM上海优秀博士生论文奖，top 2 in Shanghai
3. 2017 复旦大学研究生学术之星 复旦博士生最高学术荣誉
4. 2016 复旦大学博士优秀毕业生
5. 2016 EMC优秀奖学金
6. 2014 "微软学者"奖学金提名奖
7. 2014 复旦大学董氏奖学金
8. 2013 复旦大学毕业生之星，复旦毕业生最高荣誉，每年仅有十位学生获得
9. 2013 光华自立奖学金学术科研奖(理科)，复旦由学生评选的最高奖项
10. 2013 SIGMOD Student Travel Award
11. 2013 上海市优秀毕业生
12. 2013 勒卡斯杯数据挖掘竞赛 第二名
13. 2012 谷歌优秀学生奖学金，为了表彰杰出学术贡献
14. 2011,2009 ACM/ICPC 北京/上海赛区邀请赛 两获金牌
15. 2009 百度之星程序设计竞赛决赛 （全国前50）

Awards 

1. "2017 ACM China Excellent Doctoral Thesis" nomination (top 4 in China)
2. "2017 ACM Shanghai Excellent Doctoral Thesis award" (top 2 in Shanghai)
3. Fudan Academic Star  (top 10 PhD students in Fudan. I'm the only one from the computer science field).
4. Fudan excellent gradudate student
5. 2016 EMC outstanding scholarship
6. 2014 Microsoft Fellowship Nomination
7. 2013 Fudan Outstanding Undergraduate Award top honor for Fudan undergraduates. 10 students win this award each year.
8. 2013 Guanghua Self-Reliance Scholarship for Academic (science), top student-organized scholarship in Fudan University
9. 2013 SIGMOD Student Travel Award
10. 2013 Shanghai Outstanding Graduate Award
11. 2013 second place in LECAST data mining competition
12. 2012 Google Excellence Scholarship, for having demonstrated superior academic achievement
13. 2011, 2009 ACM/ICPC Beijing/Shanghai Region Invitation Contest double gold medals
14. 2009 BAIDU ASTAR Programming Contest Top 50