Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userResponsible Disclosure of Generative Models Using Scalable Fingerprinting
290
0
0.0
(
0
)
Added by
Ning Yu
Publication date
2020
fields
Informatics Engineering
and research's language is
English
Ask ChatGPT about the research
No Arabic abstract
Over the past six years, deep generative models have achieved a qualitatively new level of performance. Generated data has become difficult, if not impossible, to be distinguished from real data. While there are plenty of use cases that benefit from this technology, there are also strong concerns on how this new technology can be misused to spoof sensors, generate deep fakes, and enable misinformation at scale. Unfortunately, current deep fake detection methods are not sustainable, as the gap between real and fake continues to close. In contrast, our work enables a responsible disclosure of such state-of-the-art generative models, that allows researchers and companies to fingerprint their models, so that the generated samples containing a fingerprint can be accurately detected and attributed to a source. Our technique achieves this by an efficient and scalable ad-hoc generation of a large population of models with distinct fingerprints. Our recommended operation point uses a 128-bit fingerprint which in principle results in more than $10^{36}$ identifiable models. Experiments show that our method fulfills key properties of a fingerprinting mechanism and achieves effectiveness in deep fake detection and attribution.
rate research
Read More
This paper presents a novel fingerprinting scheme for the Intellectual Property (IP) protection of Generative Adversarial Networks (GANs). Prior solutions for classification models adopt adversarial examples as the fingerprints, which can raise stealthiness and robustness problems when they are applied to the GAN models. Our scheme constructs a composite deep learning model from the target GAN and a classifier. Then we generate stealthy fingerprint samples from this composite model, and register them to the classifier for effective ownership verification. This scheme inspires three concrete methodologies to practically protect the modern GAN models. Theoretical analysis proves that these methods can satisfy different security requirements necessary for IP protection. We also conduct extensive experiments to show that our solutions outperform existing strategies in terms of stealthiness, functionality-preserving and unremovability.
A promising class of generative models maps points from a simple distribution to a complex distribution through an invertible neural network. Likelihood-based training of these models requires restricting their architectures to allow cheap computation of Jacobian determinants. Alternatively, the Jacobian trace can be used if the transformation is specified by an ordinary differential equation. In this paper, we use Hutchinsons trace estimator to give a scalable unbiased estimate of the log-density. The result is a continuous-time invertible generative model with unbiased density estimation and one-pass sampling, while allowing unrestricted neural network architectures. We demonstrate our approach on high-dimensional density estimation, image generation, and variational inference, achieving the state-of-the-art among exact likelihood methods with efficient sampling.
Physical layer authentication relies on detecting unique imperfections in signals transmitted by radio devices to isolate their fingerprint. Recently, deep learning-based authenticators have increasingly been proposed to classify devices using these fingerprints, as they achieve higher accuracies compared to traditional approaches. However, it has been shown in other domains that adding carefully crafted perturbations to legitimate inputs can fool such classifiers. This can undermine the security provided by the authenticator. Unlike adversarial attacks applied in other domains, an adversary has no control over the propagation environment. Therefore, to investigate the severity of this type of attack in wireless communications, we consider an unauthorized transmitter attempting to have its signals classified as authorized by a deep learning-based authenticator. We demonstrate a reinforcement learning-based attack where the impersonator--using only the authenticators binary authentication decision--distorts its signals in order to penetrate the system. Extensive simulations and experiments on a software-defined radio testbed indicate that at appropriate channel conditions and bounded by a maximum distortion level, it is possible to fool the authenticator reliably at more than 90% success rate.
This paper proposes a new approach to detecting neural Trojans on Deep Neural Networks during inference. This approach is based on monitoring the inference of a machine learning model, computing the attribution of the models decision on different features of the input, and then statistically analyzing these attributions to detect whether an input sample contains the Trojan trigger. The anomalous attributions, aka misattributions, are then accompanied by reverse-engineering of the trigger to evaluate whether the input sample is truly poisoned with a Trojan trigger. We evaluate our approach on several benchmarks, including models trained on MNIST, Fashion MNIST, and German Traffic Sign Recognition Benchmark, and demonstrate the state of the art detection accuracy.
RF devices can be identified by unique imperfections embedded in the signals they transmit called RF fingerprints. The closed set classification of such devices, where the identification must be made among an authorized set of transmitters, has been well explored. However, the much more difficult problem of open set classification, where the classifier needs to reject unauthorized transmitters while recognizing authorized transmitters, has only been recently visited. So far, efforts at open set classification have largely relied on the utilization of signal samples captured from a known set of unauthorized transmitters to aid the classifier learn unauthorized transmitter fingerprints. Since acquiring new transmitters to use as known transmitters is highly expensive, we propose to use generative deep learning methods to emulate unauthorized signal samples for the augmentation of training datasets. We develop two different data augmentation techniques, one that exploits a limited number of known unauthorized transmitters and the other that does not require any unauthorized transmitters. Experiments conducted on a dataset captured from a WiFi testbed indicate that data augmentation allows for significant increases in open set classification accuracy, especially when the authorized set is small.
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
