Semi-supervised (SS) inference has received much attention in recent years. Apart from a moderate-sized labeled data, $\mathcal L$, the SS setting is characterized by an additional, much larger sized, unlabeled data, $\mathcal U$. The setting of $|\mathcal U\ |\gg |\mathcal L\ |$, makes SS inference unique and different from the standard missing data problems, owing to natural violation of the so-called ‘positivity’ or ‘overlap’ assumption. However, most of the SS literature implicitly assumes $\mathcal L$ and $\mathcal U$ to be equally distributed, i.e., no selection bias in the labeling. Inferential challenges in missing at random type labeling allowing for selection bias, are inevitably exacerbated by the decaying nature of the propensity score (PS). We address this gap for a prototype problem, the estimation of the response’s mean. We propose a double robust SS mean estimator and give a complete characterization of its asymptotic properties. The proposed estimator is consistent as long as either the outcome or the PS model is correctly specified. When both models are correctly specified, we provide inference results with a non-standard consistency rate that depends on the smaller size $|\mathcal L\ |$. The results are also extended to causal inference with imbalanced treatment groups. Further, we provide several novel choices of models and estimators of the decaying PS, including a novel offset logistic model and a stratified labeling model. We present their properties under both high- and low-dimensional settings. These may be of independent interest. Lastly, we present extensive simulations and also a real data application.
Note: When clicking on a Digital Object Identifier (DOI) number, you will be taken to an external site maintained by the publisher.
Some full text articles may not yet be available without a charge during the embargo (administrative interval).
What is a DOI Number?
Some links on this page may take you to non-federal websites. Their policies may differ from this site.
-
Abstract -
Abstract Global agricultural trade creates multiple telecoupled flows of nitrogen (N) and phosphorus (P). The flows of physical and virtual nutrients along with trade have discrepant effects on natural resources in different countries. However, existing literature has not quantified or analyzed such effects yet. Here we quantified the physical and virtual N and P flows embedded in the global agricultural trade networks from 1997 to 2016 and elaborated components of the telecoupling framework. The N and P flows both increased continuously and more than 25% of global consumption of nutrients in agricultural products were related to physical nutrient flows, while virtual nutrient flows were equivalent to one-third of the nutrients inputs into global agricultural system. These flows have positive telecoupling effects on saving N and P resources at the global scale. Reducing inefficient trade flows will enhance resource conservation, environmental sustainability in the hyper-globalized world.
-
Summary A fundamental challenge in semi-supervised learning lies in the observed data’s disproportional size when compared with the size of the data collected with missing outcomes. An implicit understanding is that the dataset with missing outcomes, being significantly larger, ought to improve estimation and inference. However, it is unclear to what extent this is correct. We illustrate one clear benefit: root-$n$ inference of the outcome’s mean is possible while only requiring a consistent estimation of the outcome, possibly at a rate slower than root $n$. This is achieved by a novel $k$-fold, cross-fitted, double robust estimator. We discuss both linear and nonlinear outcomes. Such an estimator is particularly suited for models that naturally do not admit root-$n$ consistency, such as high-dimensional, nonparametric or semiparametric models. We apply our methods to estimating heterogeneous treatment effects.more » « less
-
Extracellular vesicles (EVs) have gained considerable attention as vital circulating biomarkers since their structure and composition resemble the originating cells. The investigation of EVs’ biochemical and biophysical properties is of great importance to map them to their parental cells and to better understand their functionalities. In this study, a novel frequency-dependent impedance measurement system has been developed to characterize EVs based on their unique dielectric properties. The system is composed of an insulator-based dielectrophoretic (iDEP) device to entrap and immobilize a cluster of vesicles followed by utilizing electrical impedance spectroscopy (EIS) to measure their impedance at a wide frequency spectrum, aiming to analyze both their membrane and cytosolic charge-dependent contents. The EIS was initially utilized to detect nano-size vesicles with different biochemical compositions, including liposomes synthesized with different lipid compositions, as well as EVs and lipoproteins with similar biophysical properties but dissimilar biochemical properties. Moreover, EVs derived from the same parental cells but treated with different culture conditions were characterized to investigate the correlation of impedance changes with biochemical properties and functionality in terms of pro-inflammatory responses. The system also showed the ability to discriminate between EVs derived from different cellular origins as well as among size-sorted EVs harbored from the same cellular origin. This proof-of-concept approach is the first step towards utilizing EIS as a label-free, non-invasive, and rapid sensor for detection and characterization of pathogenic EVs and other nanovesicles in the future.more » « less
-
Abstract Tungsten‐doped titanium‐dioxide (W‐TiO2) nanoparticles are successfully synthesized using a multiple‐diffusion‐flame burner with a separate center tube. Vaporized titanium tetra‐isopropoxide (TTIP) precursor issues from a center tube to produce TiO2nanoparticles, while a tungsten mesh, suspended above the surrounding multiple over‐ventilated hydrogen diffusion flames, serves as the solid‐phase metal doping source. At a lower tungsten loading rate, W‐TiO2nanoparticles are generated, as indicated by an obvious angle shift of 0.15° for the entire x‐ray diffraction spectrum. However, at a higher tungsten loading rate, homogenous nucleation of WO
x occurs before or concurrently with TiO2nucleation, producing mixed nanopowders, permitting fewer tungsten ions to be doped into TiO2. Ultraviolet–visible spectroscopic characterization reveals that the as‐synthesized W‐TiO2nanoparticles possess augmented absorbing ability in the visible‐light wavelength range, where the band gap is reduced from 3.20 to 3.05 eV, compared with that for the nondoped TiO2nanoparticles.