An Adversarial Attack System for Face Recognition

Wang, Yuetian; Zhang, Chuanjing; Liao, Xuxin; Wang, Xingang; Gu, Zhaoquan

doi:10.32604/jai.2021.014175

Cited by 12 publications

(1 citation statement)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We do these two phases of optimization since empirically, we have observed more stable behavior and a better final objective value doing two rounds of optimization (compared to optimizing all of X, Y , and Z , including Z .,(cc) from the start). This becomes imperative since Z .,(cc) will play a more pronounced role in the remainder of eSVD-DE compared to all the other covariates in Z , because is it part of the “signal” that we wish to estimate and is not a “confounder.” This is inspired by theoretical results regarding warm-starting the non-convex optimization [55, 56]. This step is collectively handled by in our codebase.…”

Section: Discussionmentioning

confidence: 99%

eSVD-DE: Cohort-wide differential expression in single-cell RNA-seq data using exponential-family embeddings

Lin,

Qiu,

Roeder

2023

Preprint

View full text Add to dashboard Cite

Single-cell RNA-sequencing (scRNA) datasets are becoming increasingly popular in clinical and cohort studies, but there is a lack of methods to investigate differentially expressed (DE) genes among such datasets with numerous individuals. While numerous methods exist to find DE genes for scRNA data from limited individuals, differential-expression testing for large cohorts of case and control individuals using scRNA data poses unique challenges due to substantial effects of human variation, i.e., individual-level confounding covariates that are difficult to account for in the presence of sparsely-observed genes. In this paper, we develop the eSVD-DE, a matrix factorization that pools information across genes and removes confounding covariate effects, followed by a novel two-sample test in mean expression between case and control individuals. In general, differential testing after dimension reduction yields an inflation of Type-1 errors. However, we overcome this by testing for differences between the case and control individuals’ posterior mean distributions via a hierarchical model. We benchmark our method’s accuracy and power increase compared to other DE methods typically repurposed for analyzing cohort-wide differential expression based on simulated data. We then use the eSVD-DE to study the cohort-wide differential expression in idiopathic pulmonary fibrosis, varying severity of ulcerative colitis, and autism spectrum disorder. Altogether, eSVD-DE proposes a novel and powerful way to test for DE genes among cohorts after performing a dimension reduction.

show abstract

Section: Discussionmentioning

confidence: 99%