skip to main content


Search for: All records

Creators/Authors contains: "Pan, Yue"

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. Abstract Background Modeling of single cell RNA-sequencing (scRNA-seq) data remains challenging due to a high percentage of zeros and data heterogeneity, so improved modeling has strong potential to benefit many downstream data analyses. The existing zero-inflated or over-dispersed models are based on aggregations at either the gene or the cell level. However, they typically lose accuracy due to a too crude aggregation at those two levels. Results We avoid the crude approximations entailed by such aggregation through proposing an independent Poisson distribution (IPD) particularly at each individual entry in the scRNA-seq data matrix. This approach naturally and intuitively models the large number of zeros as matrix entries with a very small Poisson parameter. The critical challenge of cell clustering is approached via a novel data representation as Departures from a simple homogeneous IPD (DIPD) to capture the per-gene-per-cell intrinsic heterogeneity generated by cell clusters. Our experiments using real data and crafted experiments show that using DIPD as a data representation for scRNA-seq data can uncover novel cell subtypes that are missed or can only be found by careful parameter tuning using conventional methods. Conclusions This new method has multiple advantages, including (1) no need for prior feature selection or manual optimization of hyperparameters; (2) flexibility to combine with and improve upon other methods, such as Seurat. Another novel contribution is the use of crafted experiments as part of the validation of our newly developed DIPD-based clustering pipeline. This new clustering pipeline is implemented in the R (CRAN) package scpoisson . 
    more » « less
    Free, publicly-accessible full text available December 1, 2024
  2. There is a growing rise of applications that need to support a library of models with diverse latency-accuracy trade-offs on a Pareto frontier, especially in the health-care domain. This work presents an end-to-end system for training and serving weight-sharing models. On the training end, we leverage recent research in creating a family of models on the latency- accuracy Pareto frontier that share weights, reducing the total number of unique parameters. On the serving (inference end), we propose a novel accelerator FastSwitch that extracts weight reuse across different models, thereby providing fast real-time switching between different models. 
    more » « less
  3. ABSTRACT

    We use the Auriga simulations to probe different satellite quenching mechanisms operating at different mass scales ($10^5 \, \mathrm{M}_\odot \lesssim M_\star \lesssim 10^{11} \, \mathrm{M}_\odot$) in Milky Way-like hosts. Our goal is to understand the origin of the satellite colour distribution and star-forming properties in both observations and simulations. We find that the satellite populations in the Auriga simulations, which was originally designed to model Milky Way-like host galaxies, resemble the populations in the Exploration of Local VolumE Satellites (ELVES) Survey and the Satellites Around Galactic Analogs (SAGA) survey in their luminosity function in the luminosity range −12 ≲ MV ≲ −15 and resemble ELVES in their quenched fraction and colour–magnitude distribution in the luminosity range −12 ≲ Mg ≲ −15. We find that satellites transition from blue colours to red colours at the luminosity range −15 ≲ Mg ≲ −12 in both the simulations and observations and we show that this shift is driven by environmental effects in the simulations. We demonstrate also that the colour distribution in both simulations and observations can be decomposed into two statistically distinct populations based on their morphological type or star-forming status that are statistically distinct. In the simulations, these two populations also have statistically distinct infall time distributions. The comparison presented here seems to indicate that this tension is resolved by the improved target selection of ELVES, but there are still tensions in understanding the colours of faint galaxies, of which ELVES appears to have a significant population of faint blue satellites not recovered in Auriga.

     
    more » « less
  4. Abstract

    Organolead halide perovskites (OHPs) have shown unprecedented potentials in optoelectronics. However, the inherent large bandgap has restrained its working wavelength within 280–800 nm, while light at other regions, e.g., near‐infrared (NIR), may cause drastic thermal heating effect that goes against the duration of OHP devices, if not properly exploited. Herein, a solution processable and large‐scale synthesis of multifunctional OHP composites containing lanthanide‐doped upconversion nanoparticles (UCNPs) is reported. Upon NIR illumination, the upconverted photons from UCNPs at 520–550 nm can be efficiently absorbed by closely surrounded OHP nanowires (NWs) and photocurrent is subsequently generated. The narrow full width at half maximum of the absorption of rare earth ions (Yb3+and Er3+) has ensured high‐selective NIR response. Lifetime characterizations have suggested that Förster resonance energy transfer with an efficiency of 28.5% should be responsible for the direct energy transfer from UCNPs to OHP NWs. The fabricated proof‐of‐concept device has showcased perfect response to NIR light at 980 and 1532 nm, which has paved new avenues for applications of such composites in remote control, distance measurement, and stealth materials.

     
    more » « less