Search for: All records

With rich visual data, such as images, becoming readily associated with items, visually-aware recommendation systems (VARS) have been widely used in different applications. Recent studies have shown that VARS are vulnerable to item-image adversarial attacks, which add human-imperceptible perturbations to the clean images associated with those items. Attacks on VARS pose new security challenges to a wide range of applications, such as e-commerce and social media, where VARS are widely used. How to secure VARS from such adversarial attacks becomes a critical problem. Currently, there is still a lack of systematic studies on how to design defense strategies against visual attacks on VARS. In this article, we attempt to fill this gap by proposing anadversarial image denoising and detectionframework to secure VARS. Our proposed method can simultaneously (1) secure VARS from adversarial attacks characterized bylocalperturbations by image denoising based onglobalvision transformers; and (2) accurately detect adversarial examples using a novel contrastive learning approach. Meanwhile, our framework is designed to be used as both a filter and a detector so that they can bejointlytrained to improve the flexibility of our defense strategy to a variety of attacks and VARS models. Our approach is uniquely tailored for VARS, addressing the distinct challenges in scenarios where adversarial attacks can differ across industries, for instance, causing misclassification in e-commerce or misrepresentation in real estate. We have conducted extensive experimental studies with two popular attack methods (FGSM and PGD). Our experimental results on two real-world datasets show that our defense strategy against visual attacks is effective and outperforms existing methods on different attacks. Moreover, our method demonstrates high accuracy in detecting adversarial examples, complementing its robustness across various types of adversarial attacks. 

This paper presents Pesto, a high-performance Byzantine Fault Tolerant (BFT) database that offers full SQL compatibility. Pesto intentionally forgoes the use of State Machine Replication (SMR); SMR-based designs offer poor performance due to the several round trips required to order transactions. Pesto, instead, allows for replicas to remain inconsistent, and only synchronizes on demand to ensure that the database remain serializable in the presence of concurrent transactions and malicious actors. On TPC-C, Pesto matches the throughput of Peloton and Postgres, two unreplicated SQL database systems, while increasing throughput by 2.3x compared to classic SMR-based BFT-architectures, and reducing latency by 2.7x to 3.9x. Pesto's leaderless design minimizes the impact of replica failures and ensures robust performance. 

This paper presents Pesto, a high-performance Byzantine Fault Tolerant (BFT) database that offers full SQL compatibility. Pesto intentionally forgoes the use of State Machine Replication (SMR); SMR-based designs offer poor performance due to the several round trips required to order transactions. Pesto, instead, allows for replicas to remain inconsistent, and only synchronizes on demand to ensure that the database remain serializable in the presence of concurrent transactions and malicious actors. On TPC-C, Pesto matches the throughput of Peloton and Postgres, two unreplicated SQL database systems, while increasing throughput by 2.3x compared to classic SMR-based BFT-architectures, and reducing latency by 2.7x to 3.9x. Pesto's leaderless design minimizes the impact of replica failures and ensures robust performance. 

Abstract Nucleation and subsequent growth of new aerosol particles in the atmosphere is a major source of cloud condensation nuclei and persistent large uncertainty in climate models. Newly formed particles need to grow rapidly to avoid scavenging by pre-existing aerosols and become relevant for the climate and air quality. In the continental atmosphere, condensation of oxygenated organic molecules is often the dominant mechanism for rapid growth. However, the huge variety of different organics present in the continental boundary layer makes it challenging to predict nanoparticle growth rates from gas-phase measurements. Moreover, recent studies have shown that growth rates of nanoparticles derived from particle size distribution measurements show surprisingly little dependency on potentially condensable vapors observed in the gas phase. Here, we show that the observed nanoparticle growth rates in the sub-10 nm size range can be predicted in the boreal forest only for springtime conditions, even with state-of-the-art mass spectrometers and particle sizing instruments. We find that, especially under warmer conditions, observed growth is slower than predicted from gas-phase condensation. We show that only a combination of simple particle-phase reaction schemes, phase separation due to non-ideal solution behavior, or particle-phase diffusion limitations can explain the observed lower growth rates. Our analysis provides first insights as to why atmospheric nanoparticle growth rates above 10 nm h⁻¹are rarely observed. Ultimately, a reduction of experimental uncertainties and improved sub-10 nm particle hygroscopicity and chemical composition measurements are needed to further investigate the occurrence of such a growth rate-limiting process. 

Expanding gray wolf (Canis lupus) populations in Europe and North America contribute to increased risks of livestock predation, which can threaten human livelihoods and lead government agencies to target wolves for lethal removal. Public wolf hunting is a highly contentious strategy for mitigating these risks, yet few empirical studies examine its effectiveness in doing so. Using difference-in-differences and structural equation modeling of data from the northwestern US between 2005 and 2021, we analyzed impacts of wolf hunting on livestock predation by wolves and government removal of wolves in the same year and with a 1-year time lag while controlling for social and environmental variables. We found that public wolf hunting had a small negative effect on livestock predation but had no effect on government lethal removal of wolves in the same or subsequent years. Our findings challenge the assumption that wolf hunting is an effective management strategy for reducing livestock predation and lethal removal.