Write a Blog >>
ICSE 2023
Sun 14 - Sat 20 May 2023 Melbourne, Australia
Wed 17 May 2023 14:00 - 14:15 at Meeting Room 104 - AI systems engineering Chair(s): Xin Peng

Despite their stunning performance, developing deep learning models from scratch is a formidable task. Therefore, it popularizes Machine-Learning-as-a-Service (MLaaS), where general users can access the trained models of MLaaS providers via Application Programming Interfaces (APIs) on a pay-per-query basis. Unfortunately, the success of MLaaS is under threat from model extraction attacks, where attackers intend to extract a local model of equivalent functionality to the target MLaaS model. However, existing studies on model extraction of text analytics APIs frequently assume adversaries have strong knowledge about the victim model, like its architecture and parameters, which hardly holds in practice. Besides, since the attacker’s and the victim’s training data can be considerably discrepant, it is non-trivial to perform efficient model extraction. In this paper, to advance the understanding of such attacks, we propose a framework, PEEP, for practical and efficient model extraction of sentiment analysis APIs with only query access. Specifically, PEEP features a learning-based scheme, which employs out-of-domain public corpora and a novel query strategy to construct proxy training data for model extraction. Besides, PEEP introduces a greedy search algorithm to settle an appropriate architecture for the extracted model. We conducted extensive experiments with two victim models across three datasets and two real-life commercial sentiment analysis APIs. Experimental results corroborate that PEEP can consistently outperform the state-of-the-art baselines in terms of effectiveness and efficiency.

Wed 17 May

Displayed time zone: Hobart change

13:45 - 15:15
13:45
15m
Talk
FedDebug: Systematic Debugging for Federated Learning Applications
Technical Track
Waris Gill Virginia Tech, Ali Anwar University of Minnesota, Muhammad Ali Gulzar Virginia Tech
14:00
15m
Talk
Practical and Efficient Model Extraction of Sentiment Analysis APIs
Technical Track
Weibin Wu Sun Yat-sen University, Jianping Zhang The Chinese University of Hong Kong, Victor Junqiu Wei The Hong Kong Polytechnic University, Xixian Chen Tencent, Zibin Zheng School of Software Engineering, Sun Yat-sen University, Irwin King The Chinese University of Hong Kong, Michael Lyu The Chinese University of Hong Kong
14:15
15m
Talk
CrossCodeBench: Benchmarking Cross-Task Generalization of Source Code Models
Technical Track
Changan Niu Software Institute, Nanjing University, Chuanyi Li Nanjing University, Vincent Ng Human Language Technology Research Institute, University of Texas at Dallas, Richardson, TX 75083-0688, Bin Luo Nanjing University
Pre-print
14:30
15m
Talk
Challenges in Adopting Artificial Intelligence Based User Input Verification Framework in Reporting Software Systems
SEIP - Software Engineering in Practice
Dong Jae Kim Concordia University, Tse-Hsun (Peter) Chen Concordia University, Steve Sporea , Andrei Toma ERA Environmental Management Solutions, Laura Weinkam , Sarah Sajedi ERA Environmental Management Solutions, Steve Sporea
14:45
7m
Talk
Towards Understanding Quality Challenges of the Federated Learning for Neural Networks: A First Look from the Lens of Robustness
Journal-First Papers
Amin Eslami Abyane University of Calgary, Derui Zhu Technical University of Munich, Roberto Souza University of Calgary, Lei Ma University of Alberta, Hadi Hemmati York University
14:52
7m
Talk
An Empirical Study of the Impact of Hyperparameter Tuning and Model Optimization on the Performance Properties of Deep Neural Networks
Journal-First Papers
Lizhi Liao Concordia University, Heng Li Polytechnique Montréal, Weiyi Shang University of Waterloo, Lei Ma University of Alberta
15:00
7m
Talk
Black-box Safety Analysis and Retraining of DNNs based on Feature Extraction and Clustering
Journal-First Papers
Mohammed Attaoui University of Luxembourg, Hazem FAHMY University of Luxembourg, Fabrizio Pastore University of Luxembourg, Lionel Briand University of Luxembourg; University of Ottawa
Link to publication Pre-print
15:07
7m
Talk
Iterative Assessment and Improvement of DNN Operational Accuracy
NIER - New Ideas and Emerging Results
Antonio Guerriero Università di Napoli Federico II, Roberto Pietrantuono Università di Napoli Federico II, Stefano Russo Università di Napoli Federico II
Pre-print