|
|
Shangxi Wu 吴尚锡 Huawei Noah's Ark Lab, Researcher kirinng0709@gmail.com |
|
北京交通大学, 计算机系 博士 2016年9月--2025年8月. 研究兴趣为可信赖人工智能,大语言模型推理 |
出版物
link to:谷歌学术主页
link to:dblp数据库引擎
|
A Disguised Wolf Is More Harmful Than a Toothless Tiger: Adaptive Malicious Code Injection Backdoor Attack Leveraging User Behavior as Triggers
(SCI-A1 IEEE T-IFS)
|
||
|
Debiasing backdoor attack: A benign application of backdoor attack in eliminating data bias
(SCI-A1 Information Science)
|
||
|
Backdoor for Debias: Mitigating Model Bias with Backdoor Attack-based Artificial Bias
(SCI-A1 IEEE TCSVT, 2024)
|
||
|
Memory as Action: Autonomous Context Curation for Long-Horizon Agentic Tasks
|
||
|
o1-coder: an o1 replication for coding
|
||
|
Are You Copying My Model? Protecting the Copyright of Large Language Models for EaaS via Backdoor Watermark
(ACL [Area Chair Award] 2023)
|
||
|
An Experimental Study of Semantic Continuity for Deep Learning Models
|
||
|
Adaptive Adversarial Logits Pairing
(ACM Trans ToMM, 2023)
|
||
|
Attention, Please! Adversarial Defense via Attention Rectification and Preservation
(ACM Trans ToMM, 2022)
|
||
|
Robust CAPTCHAs towards Malicious OCR
(IEEE Trans MM, 2020)
|
||
|
Butterfly Detection and Classification Based on Integrated YOLO Algorithm
(ICGEC, 2019)
|
||
|
基于注意力集中机制防御对抗样本的图像分类方法
(国防发明专利)
|
||
|
基于对抗样本的防御字符识别系统验证码生成方法
(发明专利)
|
研究经历
|
华为诺亚方舟实验室
|
||
|
微软亚研院(MSRA)
|
||
|
深圳鹏城实验室
|
||
|
北京交通大学交通数据分析实验室
|
||
|
国家级大学生创新创业项目:从解释性/可视化角度分析图像对抗样本
|
||
|
北京交通大学网络与智能研究所
|
||
|
蝴蝶识别大赛, CCDM 主办
|
技能
|
编程语言: Python3(Proficient), C, OpenCV, Latex |
获奖情况
|
国家研究生奖学金, 2020年 |