University of Massachusetts Amherst

About: University of Massachusetts Amherst is a education organization based out in Amherst Center, Massachusetts, United States. It is known for research contribution in the topics: Population & Galaxy. The organization has 37274 authors who have published 83965 publications receiving 3834996 citations. The organization is also known as: UMass Amherst & Massachusetts State College.

Comprehensive Privacy Analysis of Deep Learning: Passive and Active White-box Inference Attacks against Centralized and Federated Learning

Abstract: Deep neural networks are susceptible to various inference attacks as they remember information about their training data. We design white-box inference attacks to perform a comprehensive privacy analysis of deep learning models. We measure the privacy leakage through parameters of fully trained models as well as the parameter updates of models during training. We design inference algorithms for both centralized and federated learning, with respect to passive and active inference attackers, and assuming different adversary prior knowledge. We evaluate our novel white-box membership inference attacks against deep learning algorithms to trace their training data records. We show that a straightforward extension of the known black-box attacks to the white-box setting (through analyzing the outputs of activation functions) is ineffective. We therefore design new algorithms tailored to the white-box setting by exploiting the privacy vulnerabilities of the stochastic gradient descent algorithm, which is the algorithm used to train deep neural networks. We investigate the reasons why deep learning models may leak information about their training data. We then show that even well-generalized models are significantly susceptible to white-box membership inference attacks, by analyzing state-of-the-art pre-trained and publicly available models for the CIFAR dataset. We also show how adversarial participants, in the federated learning setting, can successfully run active membership inference attacks against other participants, even when the global model achieves high prediction accuracies.

PUF Modeling Attacks on Simulated and Silicon Data

Abstract: We discuss numerical modeling attacks on several proposed strong physical unclonable functions (PUFs). Given a set of challenge-response pairs (CRPs) of a Strong PUF, the goal of our attacks is to construct a computer algorithm which behaves indistinguishably from the original PUF on almost all CRPs. If successful, this algorithm can subsequently impersonate the Strong PUF, and can be cloned and distributed arbitrarily. It breaks the security of any applications that rest on the Strong PUF's unpredictability and physical unclonability. Our method is less relevant for other PUF types such as Weak PUFs. The Strong PUFs that we could attack successfully include standard Arbiter PUFs of essentially arbitrary sizes, and XOR Arbiter PUFs, Lightweight Secure PUFs, and Feed-Forward Arbiter PUFs up to certain sizes and complexities. We also investigate the hardness of certain Ring Oscillator PUF architectures in typical Strong PUF applications. Our attacks are based upon various machine learning techniques, including a specially tailored variant of logistic regression and evolution strategies. Our results are mostly obtained on CRPs from numerical simulations that use established digital models of the respective PUFs. For a subset of the considered PUFs-namely standard Arbiter PUFs and XOR Arbiter PUFs-we also lead proofs of concept on silicon data from both FPGAs and ASICs. Over four million silicon CRPs are used in this process. The performance on silicon CRPs is very close to simulated CRPs, confirming a conjecture from earlier versions of this work. Our findings lead to new design requirements for secure electrical Strong PUFs, and will be useful to PUF designers and attackers alike.

Multi-Phase Galaxy Formation: High Velocity Clouds and the Missing Baryon Problem

Abstract: The standard treatment of cooling in Cold Dark Matter halos assumes that all of the gas within a ``cooling radius'' cools and contracts monolithically to fuel galaxy formation. Here we take into account the expectation that the hot gas in galactic halos is thermally unstable and prone to fragmentation during cooling and show that the implications are more far-reaching than previously expected: allowing multi-phase cooling fundamentally alters expectations about gas infall in halos and naturally explains the bright-end cutoff in the galaxy luminosity function. We argue that cooling should proceed via the formation of high-density, 10^4 K clouds, pressure-confined within a hot gas background. The background medium has a low density, and can survive as a stable corona with a long cooling time. The fraction of baryons contained in the residual hot core grows with halo mass because the cooling density increases, and this leads to an upper-mass limit in quiescent, non-merged galaxies of ~10^11 Msun. In this scenario, galaxy formation is fueled by the infall of pressure-supported clouds. For Milky-Way-size systems, clouds of mass ~ 5x10^6 Msun that formed or merged within the last several Gyrs should still exist as a residual population in the halo, with a total mass in clouds of ~ 2 x 10^10 Msun. The mass of the Milky Way galaxy is explained naturally in this model, and is a factor of two smaller than would result in the standard treatment without feedback. We expect clouds in galactic halos to be ~ 1 kpc in size and to extend ~150 kpc from galactic centers. The predicted properties of clouds match well the observed radial velocities, angular sizes, column densities, and velocity widths of High Velocity Clouds around our Galaxy. The clouds also explain high-ion absorption systems at z<1.