Qiongkai Xu

Qiongkai Xu

I am a research fellow on Security in NLP at the University of Melbourne (UoM). I have finished my Ph.D. programme at the Australian National University (ANU). Before that, I graduated from Shanghai Jiao Tong University (SJTU). I have worked several years for industry labs, Huawei Noah’s Ark Lab, Data61 CSIRO, IBM China Research Lab, and Microsoft Research Asia (intern). My research generally lies in among Privacy & Security, Natural Language Processing and Machine Learning. Recently, I am interested in auditing machine learning models, e.g., 1) privacy and security issues in ML/NLP models and 2) new evaluation paradigms for ML/NLP models.

If you are intrested in working with me, on some common research goals, please feel free to contact me.

Email: qiongkai.xu[at]unimelb.edu.au

Recent News

[Jan, 2023] One paper is accepted to ICLR’23 (spotlight). If you would like to certify a ‘superhuman’ machine learning model, check our latest paper.

[Jan, 2023] Congrats to Yujin and Terry! One paper ‘Training-Free Lexical Backdoor attacks on Language Models’ is accepted to WWW’23 (acceptance rate 19.2%).

[Oct, 2022] Two papers are accepted to EMNLP main conference. Congrats to Zhuang and Xuanli!

[Sep, 2022] One paper is accepted to NeurIPS. Check our work on conditional watermarks for NLP APIs.

[Sep, 2022] I am honored to be invited to give a talk at TrustML YSS on ‘Humanly Certifying Superhuman Classifiers’.

[Aug, 2022] One paper is accepted to COLING as oral presentation.

[Jun, 2022] I will join NLP Group @ University of Melbourne as a Research Fellow on Security in NLP.


Recent and Selected Publications

A complete list of publications: [Google Scholar]

On the Certification of Classifiers for Outperforming Human Annotators
ArXiv version: Humanly Certifying Superhuman Classifiers
Qiongkai Xu, Christian Walder, Chenchen Xu
(To appear at ICLR 2023, spotlight).

Training-Free Lexical Backdoor attacks on Language Models
Yujin Huang, Terry Zhuo Yue, Qiongkai Xu, Han Hu, Xingliang Yuan, Chunyang Chen
(Accepted to WWW 2023).

Variational Autoencoder with Disentanglement Priors for Low-Resource Task-Specific Natural Language Generation
Zhuang Li, Lizhen Qu, Qiongkai Xu, Tongtong Wu, Tianyang Zhan, Gholamreza Haffari
Proceedings of The 2022 Conference on Empirical Methods in Natural Language Processing (EMNLP), Dec 2022.

Extracted BERT Model Leaks More Information than You Think!
Xuanli He, Chen Chen, Lingjuan Lyu, Qiongkai Xu
Proceedings of The 2022 Conference on Empirical Methods in Natural Language Processing (EMNLP), Dec 2022.

CATER: Intellectual Property Protection on Text Generation APIs via Conditional Watermarks
Xuanli He, Qiongkai Xu, Yi Zeng, Lingjuan Lyu, Fangzhao Wu, Jiwei Li, Ruoxi Jia
Proceedings of The 36th Conference on Neural Information Processing Systems (NeurIPS), Nov 2022.

Student Surpasses Teacher: Imitation Attack for Black-Box NLP APIs
Qiongkai Xu, Xuanli He, Lingjuan Lyu, Lizhen Qu, Gholamreza Haffari
Proceedings of The 29th International Conference on Computational Linguistics (COLING), Oct 2022.

Protecting intellectual property of language generation apis with lexical watermark
Xuanli He, Qiongkai Xu, Lingjuan Lyu, Fangzhao Wu, Chenguang Wang
Proceedings of The AAAI Conference on Artificial Intelligence (AAAI), Feb 2022.

Humanly Certifying Superhuman Classifiers
Qiongkai Xu, Christian Walder, Chenchen Xu
Pre-print, Sep 2021.

Personal Information Leakage Detection in Conversations
Qiongkai Xu, Lizhen Qu, Zeyu Gao, Gholamreza Haffari
Proceedings of The 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), Nov 2020.

Adhering, Steering, and Queering: Treatment of Gender in Natural Language Generation
Yolande Strengers, Lizhen Qu, Qiongkai Xu, Jarrod Knibbe
Proceedings of The 2020 CHI Conference on Human Factors in Computing Systems (CHI), Apr 2020.

Privacy-Aware Text Rewriting
Qiongkai Xu, Lizhen Qu, Chenchen Xu, Ran Cui
Proceedings of The 12th International Conference on Natural Language Generation (INLG), Oct 2019.

Deep Neural networks for Learning Graph Representations
Shaosheng Cao, Wei Lu, Qiongkai Xu
Proceedings of The AAAI Conference on Artificial Intelligence (AAAI), Feb 2016.

GraRep: Learning Graph Representations with Global Structural Information
Shaosheng Cao, Wei Lu, Qiongkai Xu
Proceedings of The 24th ACM international on conference on information and knowledge management (CIKM), Oct 2015.


Misc

I boulder regularly and aim at climbing high someday. I believe that doing research is similar to bouldering and climbing.

This website is still under construction.