skip to main content

Explore Research Products in the NSF-PAR

This content will become publicly available on July 1, 2024

Title: Efficient median of means estimator
The goal of this note is to present a modification of the popular median of means estimator that achieves sub-Gaussian deviation bounds with nearly optimal constants under minimal assumptions on the underlying distribution. We build on the recent work on the topic and prove that desired guarantees can be attained under weaker requirements.  more » « less
Award ID(s):
2045068 1908905
NSF-PAR ID:
10423178
Author(s) / Creator(s):
Date Published:
Journal Name:
Proceedings of Machine Learning Research
Volume:
195
ISSN:
2640-3498
Page Range / eLocation ID:
1-9
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. ; ; ; ; ( , Materials characterization)
    null (Ed.)
    Stress-strain responses and twinning characteristics are studied for a rolled AZ31B magnesium alloy under three different stress states: tension along the normal direction (NDT), compression along the rolled direction (RDC), and torsion about the normal direction (NDTOR) using companion specimens interrupted at incremental strain levels. Tension twinning is extensively induced in twinning-favorable NDT and RDC. All the six variants of tension twin are activated under NDT, whereas a maximum of four variants is activated under RDC. Under NDTOR, both tension twins and compression twins are activated at relatively large strains and twinning occurs in a small fraction of favored grains rather than in the majority of grains. Secondary and tertiary twins are observed in the favorably-orientated grains at high strain levels. Deformation under each stress state shows three stages of strain hardening rate: fast decrease (Stage I), sequential increase (Stage II), and progressive decrease (Stage III). The increase in the hardening rate, which is more significant under NDT and RDC as compared to NDTOR, is attributed to the hardening effect of twin boundaries and twinning texture-induced slip activities. The hardening effect of twin boundaries include the dynamic Hall-Petch hardening induced by the multiplication of twin boundaries (TBs) and twin-twin boundaries (TTBs) as well as the hardening effect associated with the energetically unfavorable TTB formation. When the applied plastic strain is larger than 0.05 under NDT and RDC, the tension twin volume fraction is higher than 50%. The twinning-induced texture leads to the activation of non-basal slips mainly in the twinned volume, i.e. prismatic slips under NDT and pyramidal slips under RDC. The low work hardening under NDTOR is due to the prevailing basal slips with reduced twinning activities under NDTOR. 
    more » « less
  2. ; ( , Journal of the Royal Statistical Society Series B: Statistical Methodology)
    Summary

    Covariate-adaptive randomization is popular in clinical trials with sequentially arrived patients for balancing treatment assignments across prognostic factors that may have influence on the response. However, existing theory on tests for the treatment effect under covariate-adaptive randomization is limited to tests under linear or generalized linear models, although the covariate-adaptive randomization method has been used in survival analysis for a long time. Often, practitioners will simply adopt a conventional test to compare two treatments, which is controversial since tests derived under simple randomization may not be valid in terms of type I error under other randomization schemes. We derive the asymptotic distribution of the partial likelihood score function under covariate-adaptive randomization and a working model that is subject to possible model misspecification. Using this general result, we prove that the partial likelihood score test that is robust against model misspecification under simple randomization is no longer robust but conservative under covariate-adaptive randomization. We also show that the unstratified log-rank test is conservative and the stratified log-rank test remains valid under covariate-adaptive randomization. We propose a modification to variance estimation in the partial likelihood score test, which leads to a score test that is valid and robust against arbitrary model misspecification under a large family of covariate-adaptive randomization schemes including simple randomization. Furthermore, we show that the modified partial likelihood score test derived under a correctly specified model is more powerful than log-rank-type tests in terms of Pitman’s asymptotic relative efficiency. Simulation studies about the type I error and power of various tests are presented under several popular randomization schemes.

     
    more » « less
  3. ; ; ; ; ; ; ; ( , Advances in Microbiology)
    The term “microgravity” is used to describe the “weightlessness” or “zero-g” circumstances that can only be found in space beyond earth’s atmosphere. Rhodobacter sphaeroides is a gram-negative purple phototroph, used as a model organism for this study due to its genomic complexity and metabolic versatility. Its genome has been completely sequenced, and profiles of the differential gene expression under aerobic, semi-aerobic, and photosynthetic conditions were examined. In this study, we hypothesized that R. sphaeroides will show altered growth characteristics, morphological properties, and gene expression patterns when grown under simulated microgravity. To test that, we measured the optical density and colony-forming units of cell cultures grown under both microgravity and normal gravity conditions. Differences in the cell morphology were observed using scanning electron microscopy (SEM) images by measuring the length and the surface area of the cells under both conditions. Furthermore, we also identified homologous genes of R. spheroides using the differential gene expression study of Acidovorax under microgravity in our laboratory. Growth kinetics results showed that R. sphaeroides cells grown under microgravity experience a shorter log phase and early stationary phase compared to the cells growing under normal gravity conditions. The length and surface area of the cells under microgravity were significantly higher confirming that bacterial cells experience altered morphological features when grown under microgravity conditions. Differentially expressed homologous gene analysis indicated that genes coding for several COG and GO functions, such as metabolism, signal-transduction, transcription, translation, chemotaxis, and cell motility are differentially expressed to adapt and survive microgravity. 
    more » « less
  4. ; ; ; ; ; ; ; ; ; ; et al ( , Global Biogeochemical Cycles)
    Abstract

    The terrestrial carbon sink provides a critical negative feedback to climate warming, yet large uncertainty exists on its long‐term dynamics. Here we combined terrestrial biosphere models (TBMs) and climate projections, together with climate‐specific land use change, to investigate both the trend and interannual variability (IAV) of the terrestrial carbon sink from 1986 to 2099 under two representative concentration pathways RCP2.6 and RCP6.0. The results reveal a saturation of the terrestrial carbon sink by the end of this century under RCP6.0 due to warming and declined CO2effects. Compared to 1986–2005 (0.96 ± 0.44 Pg C yr−1), during 2080–2099 the terrestrial carbon sink would decrease to 0.60 ± 0.71 Pg C yr−1but increase to 3.36 ± 0.77 Pg C yr−1, respectively, under RCP2.6 and RCP6.0. The carbon sink caused by CO2, land use change and climate change during 2080–2099 is −0.08 ± 0.11 Pg C yr−1, 0.44 ± 0.05 Pg C yr−1, and 0.24 ± 0.70 Pg C yr−1under RCP2.6, and 4.61 ± 0.17 Pg C yr−1, 0.22 ± 0.07 Pg C yr−1, and ‐1.47 ± 0.72 Pg C yr−1under RCP6.0. In addition, the carbon sink IAV shows stronger variance under RCP6.0 than RCP2.6. Under RCP2.6, temperature shows higher correlation with the carbon sink IAV than precipitation in most time, which however is the opposite under RCP6.0. These results suggest that the role of terrestrial carbon sink in curbing climate warming would be weakened in a no‐mitigation world in future, and active mitigation efforts are required as assumed under RCP2.6.

     
    more » « less
  5. ; ; ; ; ; ; ; ; ; ; et al ( , Elementa: Science of the Anthropocene)
    During the Arctic melt season, relatively fresh meltwater layers can accumulate under sea ice as a result of snow and ice melt, far from terrestrial freshwater inputs. Such under-ice meltwater layers, sometimes referred to as under-ice melt ponds, have been suggested to play a role in the summer sea ice mass balance both by isolating the sea ice from saltier water below, and by driving formation of ‘false bottoms’ below the sea ice. Such layers form at the interface of the fresher under-ice layer and the colder, saltier seawater below. During the Multidisciplinary drifting Observatory for the Study of the Arctic Climate (MOSAiC) expedition in the Central Arctic, we observed the presence of under-ice meltwater layers and false bottoms throughout July 2020 at primarily first-year ice locations. Here, we examine the distribution, prevalence, and drivers of under-ice ponds and the resulting false bottoms during this period. The average thickness of observed false bottoms and freshwater equivalent of under-ice meltwater layers was 0.08 m, with false bottom ice comprised of 74–87% FYI melt and 13–26% snow melt. Additionally, we explore these results using a 1D model to understand the role of dynamic influences on decoupling the ice from the seawater below. The model comparison suggests that the ice-ocean friction velocity was likely exceptionally low, with implications for air-ice-ocean momentum transfer. Overall, the prevalence of false bottoms was similar to or higher than noted during other observational campaigns, indicating that these features may in fact be common in the Arctic during the melt season. These results have implications for the broader ice-ocean system, as under-ice meltwater layers and false bottoms provide a source of ice growth during the melt season, potentially reduce fluxes between the ice and the ocean, isolate sea ice primary producers from pelagic nutrient sources, and may alter light transmission to the ocean below. 
    more » « less
  • Citation Formats
  • MLA
  • APA
  • Chicago
  • BibTeX
  • Save / Share this Record
  • Send to Email