Exploring Connections Between Active Learning and Model Extraction

Authors: 

Varun Chandrasekaran, University of Wisconsin-Madison; Kamalika Chaudhuri, University of California San Diego; Irene Giacomelli, Protocol Labs; Somesh Jha, University of Wisconsin-Madison; Songbai Yan, University of California San Diego

Abstract: 

Machine learning is being increasingly used by individuals, research institutions, and corporations. This has resulted in the surge of Machine Learning-as-a-Service (MLaaS) - cloud services that provide (a) tools and resources to learn the model, and (b) a user-friendly query interface to access the model. However, such MLaaS systems raise privacy concerns such as \emph{model extraction}. In model extraction attacks, adversaries maliciously exploit the query interface to {\em steal} the model. More precisely, in a model extraction attack, a good approximation of a sensitive or proprietary model held by the server is extracted (\ie learned) by a dishonest user who interacts with the server only via the query interface. This attack was introduced by Tram`{e}r \etal at the 2016 USENIX Security Symposium, where practical attacks for various models were shown. We believe that better understanding the efficacy of model extraction attacks is paramount to designing secure MLaaS systems. To that end, we take the first step by (a) formalizing model extraction and discussing possible defense strategies, and (b) drawing parallels between model extraction and established area of \emph{active learning}. In particular, we show that recent advancements in the active learning domain can be used to implement powerful model extraction attacks and investigate possible defense strategies.

Open Access Media

USENIX is committed to Open Access to the research presented at our events. Papers and proceedings are freely available to everyone once the event begins. Any video, audio, and/or slides that are posted after the event are also free and open to everyone. Support USENIX and our commitment to Open Access.

BibTeX
@inproceedings {247638,
title = {Exploring Connections Between Active Learning and Model Extraction},
booktitle = {29th {USENIX} Security Symposium ({USENIX} Security 20)},
year = {2020},
address = {Boston, MA},
url = {https://www.usenix.org/conference/usenixsecurity20/presentation/chandrasekaran},
publisher = {{USENIX} Association},
month = aug,
}