Results 81 to 90 of about 82,924 (315)

On Second Thought: The Impact of Confessions, DNA, and Belief Perseverance on Students' Perceptions of Guilt and Interrogations

Behavioral Sciences &the Law, EarlyView.
ABSTRACT Despite growing public knowledge of false confession cases, research with students and community members continues to find that people assume confessions indicate guilt. The present research explored the implications of belief perseverance: the tendency to maintain a belief even when confronted with compelling contradictory evidence.
Taya D. Henry, Kimberley A. Clow, Victoria Hall   +2 more
wiley   +1 more source

Adversarial defenses via vector quantization

Neurocomputing
This is the author-accepted version of our paper published in Neurocomputing.
Zhiyi Dong, Yongyi Mao
openaire   +2 more sources

Applying the Rules of Evidence to Expert Testimony About Risk

Behavioral Sciences &the Law, EarlyView.
ABSTRACT Expert opinion about dangerousness or risk is common at sentencing, criminal commitment proceedings and some types of pretrial detention hearings. This article argues that such evidence must be (1) “material” (logically relevant, empirically generalizable, and epistemologically germane), (2) “probative” (a measure of accuracy, which is ...
Christopher Slobogin
wiley   +1 more source

A Review of Overcurrent Protection in Smart Grids Under Cyber‐Physical Threats With a Cyber‐Physical Evaluation Framework

Energy Science &Engineering, EarlyView.
By manipulating current and voltage measurements, an assailant can induce unwanted relay action while attempting to avoid detection. Detecting advanced cyber intrusions in power protection environments requires specialised data analysis and anomaly detection methods.
Feras Alasali, Anas AlMajali, Naser Naily, Hassen Loukil, Salah AbuGhalyon, William Holderbaum, Abdelaziz Salah Saidi   +6 more
wiley   +1 more source

A Deep Learning Framework for Forecasting Medium‐Term Covariance in Multiasset Portfolios

Journal of Forecasting, EarlyView.
ABSTRACT Forecasting the covariance matrix of asset returns is central to portfolio construction, risk management, and asset pricing. However, most existing models struggle at medium‐term horizons, several weeks to months, where shifting market regimes and slower dynamics prevail.
Pedro Reis, Ana Paula Serra, João Gama
wiley   +1 more source

A knowledge distillation strategy for enhancing the adversarial robustness of lightweight automatic modulation classification models

IET Communications
Automatic modulation classification models based on deep learning models are at risk of being interfered by adversarial attacks. In an adversarial attack, the attacker causes the classification model to misclassify the received signal by adding carefully
Fanghao Xu, Chao Wang, Jiakai Liang, Chenyang Zuo, Keqiang Yue, Wenjun Li   +5 more
doaj   +1 more source

PuVAE: A Variational Autoencoder to Purify Adversarial Examples

IEEE Access, 2019
Deep neural networks are widely used and exhibit excellent performance in many areas. However, they are vulnerable to adversarial attacks that compromise networks at inference time by applying elaborately designed perturbations to input data.
Uiwon Hwang, Jaewoo Park, Hyemi Jang, Sungroh Yoon, Nam Ik Cho   +4 more
doaj   +1 more source

Adversarial Traces for Website Fingerprinting Defense [PDF]

Proceedings of the 2018 ACM SIGSAC Conference on Computer and Communications Security, 2018
Website Fingerprinting (WF) is a traffic analysis attack that enables an eavesdropper to infer the victim's web activity even when encrypted and even when using the Tor anonymity system. Using deep learning classifiers, the attack can reach up to 98% accuracy.
Mohsen Imani, Mohammad Saidur Rahman 0002, Matthew Wright 0001   +2 more
openaire   +1 more source

Knowledge sourcing, geopolitics, and FDI: An empirical analysis on the US green and digital sectors

Global Strategy Journal, EarlyView.
Abstract Research Summary This paper examines how foreign direct investment (FDI) shapes firms' sourcing of knowledge in the digital and green domains under rising geopolitical frictions. We assemble a firm–country dyadic panel (2013–2020) linking US patent backward citations to firms' FDI, enriched with bilateral geopolitical distance and host‐country
Alberto Maria Radici
wiley   +1 more source

Attention, Please! Adversarial Defense via Attention Rectification and Preservation

, 2019
This study provides a new understanding of the adversarial attack problem by examining the correlation between adversarial attack and visual attention change.
Jing, Liping, Sang, Jitao, Sun, Yanfeng, Wu, Shangxi, Xu, Kaiyuan, Yu, Jian, Zhang, Jiaming   +6 more
core