Backdoors in Machine Learning Models
Clean-Label Attacks
One disadvantage of the approach described in this article is that the manipulations are easy to detect. First, the trigger could be found in the training examples. Second, the training examples with triggers have an incorrect label, the one that the attacker wants the model to provide as a response when the trigger is present. Some more advanced approaches might try to hide the manipulations. In clean-label attacks, only the image data is manipulated. The labels remain unchanged so that the label still matches the image. And the image data can even be manipulated in a way that it is imperceptible to the people reviewing the data set.
To inject a backdoor into a model, you do not necessarily need to manipulate an existing data set or create a new, manipulated, labeled data set. Instead, all you need to do is post manipulated images at certain places on the Internet, where they will presumably be accessed by someone at some point, in order to create a model from them. In this case, the images would be labeled by other people (for example, via crowdsourcing) who would not notice the manipulations.
Conclusions
Machine learning and smart systems are currently making giant inroads into every area of daily life. The potential is enormous, and impressive results are repeatedly achieved. But progress always goes hand in hand with new risks. Although the security properties of machine learning models have now been far more thoroughly investigated than a few years ago, still very little is known about them. The AI community will need to develop more effective protections against data poisoning attacks before we can truly trust our smart systems.
Infos
- Szegedy, Christian, et al. "Intriguing properties of neural networks." arXiv:1312.6199, Dec. 2013
- Athalye, Anish, Logan Engstrom, Andrew Ilyas, and Kevin Kwok. "Synthesizing robust adversarial examples." Proceedings of the 35th International Conference on Machine Learning (2018), PMLR 80:284-293
- Goodfellow, Ian J., Jonathon Shlens, and Christian Szegedy. "Explaining and harnessing adversarial examples." arXiv:1412.6572 [stat.ML], Dec. 2014
- Fredrikson, Matt, Somesh Jha, and Thomas Ristenpart. "Model inversion attacks that exploit confidence information and basic countermeasures." Proceedings of the 22nd ACM SIGSAC Conference on Computer and Communications Security (2015), pg. 1322-1333
- Amazon's Mechanical Turk: https://www.mturk.com
- Tianyu Gu, Brendan Dolan-Gavitt, and Siddharth Garg. "Badnets: Identifying vulnerabilities in the machine learning model supply chain." arXiv:1708.06733 [cs.CR], Aug. 2017
- Deng, L. "The MNIST Database of Handwritten Digit Images for Machine Learning Research." IEEE Signal Processing Magazine, 2012;29(6):141-142
- Jupyter Notebook: https://github.com/daniel-e/secml/blob/master/examples/backdoors/mnist.ipynb
« Previous 1 2 3 4
Buy this article as PDF
(incl. VAT)
Buy Linux Magazine
Subscribe to our Linux Newsletters
Find Linux and Open Source Jobs
Subscribe to our ADMIN Newsletters
Support Our Work
Linux Magazine content is made possible with support from readers like you. Please consider contributing when you’ve found an article to be beneficial.
News
-
Red Hat Adds New Deployment Option for Enterprise Linux Platforms
Red Hat has re-imagined enterprise Linux for an AI future with Image Mode.
-
OSJH and LPI Release 2024 Open Source Pros Job Survey Results
See what open source professionals look for in a new role.
-
Proton 9.0-1 Released to Improve Gaming with Steam
The latest release of Proton 9 adds several improvements and fixes an issue that has been problematic for Linux users.
-
So Long Neofetch and Thanks for the Info
Today is a day that every Linux user who enjoys bragging about their system(s) will mourn, as Neofetch has come to an end.
-
Ubuntu 24.04 Comes with a “Flaw"
If you're thinking you might want to upgrade from your current Ubuntu release to the latest, there's something you might want to consider before doing so.
-
Canonical Releases Ubuntu 24.04
After a brief pause because of the XZ vulnerability, Ubuntu 24.04 is now available for install.
-
Linux Servers Targeted by Akira Ransomware
A group of bad actors who have already extorted $42 million have their sights set on the Linux platform.
-
TUXEDO Computers Unveils Linux Laptop Featuring AMD Ryzen CPU
This latest release is the first laptop to include the new CPU from Ryzen and Linux preinstalled.
-
XZ Gets the All-Clear
The back door xz vulnerability has been officially reverted for Fedora 40 and versions 38 and 39 were never affected.
-
Canonical Collaborates with Qualcomm on New Venture
This new joint effort is geared toward bringing Ubuntu and Ubuntu Core to Qualcomm-powered devices.