skip to main content


Search for: All records

Creators/Authors contains: "Ma, L."

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.

  1. Summary Identifying dependency in multivariate data is a common inference task that arises in numerous applications. However, existing nonparametric independence tests typically require computation that scales at least quadratically with the sample size, making it difficult to apply them in the presence of massive sample sizes. Moreover, resampling is usually necessary to evaluate the statistical significance of the resulting test statistics at finite sample sizes, further worsening the computational burden. We introduce a scalable, resampling-free approach to testing the independence between two random vectors by breaking down the task into simple univariate tests of independence on a collection of $2\times 2$ contingency tables constructed through sequential coarse-to-fine discretization of the sample , transforming the inference task into a multiple testing problem that can be completed with almost linear complexity with respect to the sample size. To address increasing dimensionality, we introduce a coarse-to-fine sequential adaptive procedure that exploits the spatial features of dependency structures. We derive a finite-sample theory that guarantees the inferential validity of our adaptive procedure at any given sample size. We show that our approach can achieve strong control of the level of the testing procedure at any sample size without resampling or asymptotic approximation and establish its large-sample consistency. We demonstrate through an extensive simulation study its substantial computational advantage in comparison to existing approaches while achieving robust statistical power under various dependency scenarios, and illustrate how its divide-and-conquer nature can be exploited to not just test independence, but to learn the nature of the underlying dependency. Finally, we demonstrate the use of our method through analysing a dataset from a flow cytometry experiment. 
    more » « less
  2. Seawater microorganisms play an important role in coral reef ecosystem functioning and can be influenced by biological, chemical, and physical features of reefs. As coral reefs continue to respond to environmental changes, the reef seawater microbiome has been proposed as a conservation tool for monitoring perturbations. However, the spatial variability of reef seawater microbial communities is not well studied, limiting our ability to make generalizable inferences across reefs. In order to better understand how microorganisms are distributed at multiple spatial scales, we examined seawater microbial communities in Florida Reef Tract and US Virgin Islands reef systems using a nested sampling design. On 3 reefs per reef system, we sampled seawater at regular spatial intervals close to the benthos. We assessed the microbial community composition of these waters using ribosomal RNA gene amplicon sequencing. Our analysis revealed that reef water microbial communities varied as a function of reef system and individual reefs, but communities did not differ within reefs and were not significantly influenced by benthic composition. For the reef system and inter-reef differences, abundant microbial taxa were found to be potentially useful indicators of environmental difference due to their high prevalence and variance. We further examined reef water microbial biogeography on a global scale using a secondary analysis of 5 studies, which revealed that microbial communities were more distinct with increasing geographic distance. These results suggest that biogeography is a distinguishing feature for reef water microbiomes, and that development of monitoring criteria may necessitate regionally specific sampling and analyses. 
    more » « less
  3. null (Ed.)
  4. null (Ed.)
  5. Abstract We study Λ-type Electromagnetically Induced Transparency (EIT) on the Rb D2 transition in a buffer-gas-free thermal vapor cell without anti-relaxation coating. Experimental data show well-resolved features due to velocity-selective optical pumping and one EIT resonance. The Zeeman splitting of the EIT line in magnetic fields up to 12 Gauss is investigated. One Zeeman component is free of the first-order shift and its second-order shift agrees well with theory. The full width at half maximum (FWHM) of this magnetic-field-insensitive EIT resonance is reduced due to Doppler narrowing, scales linearly in Rabi frequency over the range studied, and reaches about 100 kHz at the lowest powers. These observations agree with an analytic model for a Doppler-broadened medium developed in (Javan et al 2002 Phys. Rev. A 66 013805; Lee et al 2003 Appl. Phys. B, Lasers Opt. (Germany) B 76 , 33–9; Taichenachev et al 2000 JETP Lett. 72 , 119). Numerical simulation using the Lindblad equation reveals that the transverse laser intensity distribution and two Λ-EIT systems must be included to fully account for the measured line width and line shape of the signals. Ground-state decoherence, caused by effects that include residual optical frequency fluctuations, atom-wall and trace-gas collisions, is discussed. 
    more » « less
  6. Abstract

    Understanding relationships between stream chemistry and watershed factors: land use/land cover, climate, and lithology are crucial to improving our knowledge of critical zone processes that influence water quality. We compiled major ion data from >100 monitoring stations collected over 60 years (1958–2018) across the Colorado River Watershed in Texas (103,000 km2). We paired this river chemistry data with complementary lithology, land use, climate, and stream discharge information. Machine learning techniques were used to produce new insights on controls of stream water chemical behavior, which were validated using traditional multivariate analyses. Studies on stream flow and chemistry in the American west and globally have shown strong relationships between major ion chemical composition, climate, and lithology which hold true for the Colorado River basin in this study. Reactive minerals, including carbonates and evaporites, dominate major ion chemistry across the upper, low‐precipitation regions of the watershed. Upstream and middle reaches of the Colorado River showed shifts from Na‐Cl‐SO4dominated water from multiple sources including dissolution of gypsum and halite in shallow groundwater, and agricultural activities, to Ca‐HCO3water types controlled by carbonate dissolution. In the lower portion of the watershed multiple analyses demonstrate that stream chemistry is more influenced by greater precipitation and the presence of silicate minerals than the middle and upstream reaches. This study demonstrates the power of applying machine learning approaches to publicly available long term water chemistry data sets to improve the understanding of watershed interactions with surficial lithology, salinity sources, and anthropogenic influences of water quality.

     
    more » « less
  7. Existing image-to-image transformation approaches primarily focus on synthesizing visually pleasing data. Generating images with correct identity labels is challenging yet much less explored. It is even more challenging to deal with image transformation tasks with large deformation in poses, viewpoints, or scales while preserving the identity, such as face rotation and object viewpoint morphing. In this paper, we aim at transforming an image with a fine-grained category to synthesize new images that preserve the identity of the input image, which can thereby benefit the subsequent fine-grained image recognition and few-shot learning tasks. The generated images, transformed with large geometric deformation, do not necessarily need to be of high visual quality but are required to maintain as much identity information as possible. To this end, we adopt a model based on generative adversarial networks to disentangle the identity related and unrelated factors of an image. In order to preserve the fine-grained contextual details of the input image during the deformable transformation, a constrained nonalignment connection method is proposed to construct learnable highways between intermediate convolution blocks in the generator. Moreover, an adaptive identity modulation mechanism is proposed to transfer the identity information into the output image effectively. Extensive experiments on the CompCars and Multi-PIE datasets demonstrate that our model preserves the identity of the generated images much better than the state-of-the-art image-to-image transformation models, and as a result significantly boosts the visual recognition performance in fine-grained few-shot learning. 
    more » « less
  8. We demonstrate laser induced DC electric fields in an all-glass vapor cell without bulk or thin film electrodes. The spatial field distribution is mapped by Rydberg electromagnetically induced transparency (EIT) spectroscopy. The fields are generated by a photoelectric effect and allow DC electric field tuning of up to 0.8 V/cm within the Rydberg EIT probe region. We explain the measured with a boundary-value electrostatic model. This work may inspire new approaches for DC electric field control in designing miniaturized atomic vapor cell devices. Limitations and other charge effects are also discussed.

     
    more » « less
  9. The need for efficiently finding the video content a user wants is increasing because of the erupting of user-generated videos on the Web. Existing keyword-based or content-based video retrieval methods usually determine what occurs in a video but not when and where. In this paper, we make an answer to the question of when and where by formulating a new task, namely spatio-temporal video re-localization. Specifically, given a query video and a reference video, spatio-temporal video re-localization aims to localize tubelets in the reference video such that the tubelets semantically correspond to the query. To accurately localize the desired tubelets in the reference video, we propose a novel warp LSTM network, which propagates the spatio-temporal information for a long period and thereby captures the corresponding long-term dependencies. Another issue for spatio-temporal video re-localization is the lack of properly labeled video datasets. Therefore, we reorganize the videos in the AVA dataset to form a new dataset for spatio-temporal video re-localization research. Extensive experimental results show that the proposed model achieves superior performances over the designed baselines on the spatio-temporal video re-localization task. 
    more » « less