arXiv subjects

Cheng Li

Publications and source records attributed to Cheng Li.

At least 325 records · Page 18Linked to original sources

arXiv2021

Evidence for multiple Ferrel-like cells on Jupiter

Jupiter's atmosphere is dominated by multiple jet streams which are strongly tied to its 3D atmospheric circulation. Lacking a rigid bottom boundary, several models exist for how the meridional circulation extends into the planetary interior. Here we show, collecting evidence from multiple instruments of the Juno mission, the existence of mid-latitudinal meridional circulation cells which are driven by turbulence, similar to the Ferrel cells on Earth. Different than Earth, which contains only one such cell in each hemisphere, the larger, faster rotating Jupiter can incorporate multiple cells. The cells form regions of upwelling and downwelling, which we show are clearly evident in Juno's microwave data between latitude 60S and 60N. The existence of these cells is confirmed by reproducing the ammonia observations using a simplistic model. This study solves a long-standing puzzle regarding the nature of Jupiter's sub-cloud dynamics and provides evidence for 8 cells in each Jovian hemisphere.

astro-ph.EP↗

arXiv2021

Wolf-Rayet galaxies in SDSS-IV MaNGA. II. Metallicity dependence of the high-mass slope of the stellar initial mass function

As hosts of living high-mass stars, Wolf-Rayet (WR) regions or WR galaxies are ideal objects for constraining the high-mass end of the stellar initial mass function (IMF). We construct a large sample of 910 WR galaxies/regions that cover a wide range of stellar metallicity (from Z~0.001 up to Z~0.03), by combining three catalogs of WR galaxies/regions previously selected from the SDSS and SDSS-IV/MaNGA surveys. We measure the equivalent widths of the WR blue bump at ~4650 A for each spectrum. They are compared with predictions from stellar evolutionary models Starburst99 and BPASS, with different IMF assumptions (high-mass slope α of the IMF ranging from 1.0 up to 3.3). Both singular evolution and binary evolution are considered. We also use a Bayesian inference code to perform full spectral fitting to WR spectra with stellar population spectra from BPASS as fitting templates. We then make model selection among different α assumptions based on Bayesian evidence. These analyses have consistently led to a positive correlation of IMF high-mass slope α with stellar metallicity Z, i.e. with steeper IMF (more bottom-heavy) at higher metallicities. Specifically, an IMF with α=1.00 is preferred at the lowest metallicity (Z~0.001), and a Salpeter or even steeper IMF is preferred at the highest metallicity (Z~0.03). These conclusions hold even when binary population models are adopted.

astro-ph.GA↗

arXiv2021

Self-Supervised Learning for MRI Reconstruction with a Parallel Network Training Framework

Image reconstruction from undersampled k-space data plays an important role in accelerating the acquisition of MR data, and a lot of deep learning-based methods have been exploited recently. Despite the achieved inspiring results, the optimization of these methods commonly relies on the fully-sampled reference data, which are time-consuming and difficult to collect. To address this issue, we propose a novel self-supervised learning method. Specifically, during model optimization, two subsets are constructed by randomly selecting part of k-space data from the undersampled data and then fed into two parallel reconstruction networks to perform information recovery. Two reconstruction losses are defined on all the scanned data points to enhance the network's capability of recovering the frequency information. Meanwhile, to constrain the learned unscanned data points of the network, a difference loss is designed to enforce consistency between the two parallel networks. In this way, the reconstruction model can be properly trained with only the undersampled data. During the model evaluation, the undersampled data are treated as the inputs and either of the two trained networks is expected to reconstruct the high-quality results. The proposed method is flexible and can be employed in any existing deep learning-based method. The effectiveness of the method is evaluated on an open brain MRI dataset. Experimental results demonstrate that the proposed self-supervised method can achieve competitive reconstruction performance compared to the corresponding supervised learning method at high acceleration rates (4 and 8). The code is publicly available at \url{https://github.com/chenhu96/Self-Supervised-MRI-Reconstruction}.

cs.CV↗

arXiv2021

Group Shift Pointwise Convolution for Volumetric Medical Image Segmentation

Recent studies have witnessed the effectiveness of 3D convolutions on segmenting volumetric medical images. Compared with the 2D counterparts, 3D convolutions can capture the spatial context in three dimensions. Nevertheless, models employing 3D convolutions introduce more trainable parameters and are more computationally complex, which may lead easily to model overfitting especially for medical applications with limited available training data. This paper aims to improve the effectiveness and efficiency of 3D convolutions by introducing a novel Group Shift Pointwise Convolution (GSP-Conv). GSP-Conv simplifies 3D convolutions into pointwise ones with 1x1x1 kernels, which dramatically reduces the number of model parameters and FLOPs (e.g. 27x fewer than 3D convolutions with 3x3x3 kernels). Naïve pointwise convolutions with limited receptive fields cannot make full use of the spatial image context. To address this problem, we propose a parameter-free operation, Group Shift (GS), which shifts the feature maps along with different spatial directions in an elegant way. With GS, pointwise convolutions can access features from different spatial locations, and the limited receptive fields of pointwise convolutions can be compensated. We evaluate the proposed methods on two datasets, PROMISE12 and BraTS18. Results show that our method, with substantially decreased model complexity, achieves comparable or even better performance than models employing 3D convolutions.

eess.IV↗

arXiv2021

A Novel Hybrid Convolutional Neural Network for Accurate Organ Segmentation in 3D Head and Neck CT Images

Radiation therapy (RT) is widely employed in the clinic for the treatment of head and neck (HaN) cancers. An essential step of RT planning is the accurate segmentation of various organs-at-risks (OARs) in HaN CT images. Nevertheless, segmenting OARs manually is time-consuming, tedious, and error-prone considering that typical HaN CT images contain tens to hundreds of slices. Automated segmentation algorithms are urgently required. Recently, convolutional neural networks (CNNs) have been extensively investigated on this task. Particularly, 3D CNNs are frequently adopted to process 3D HaN CT images. There are two issues with naïve 3D CNNs. First, the depth resolution of 3D CT images is usually several times lower than the in-plane resolution. Direct employment of 3D CNNs without distinguishing this difference can lead to the extraction of distorted image features and influence the final segmentation performance. Second, a severe class imbalance problem exists, and large organs can be orders of times larger than small organs. It is difficult to simultaneously achieve accurate segmentation for all the organs. To address these issues, we propose a novel hybrid CNN that fuses 2D and 3D convolutions to combat the different spatial resolutions and extract effective edge and semantic features from 3D HaN CT images. To accommodate large and small organs, our final model, named OrganNet2.5D, consists of only two instead of the classic four downsampling operations, and hybrid dilated convolutions are introduced to maintain the respective field. Experiments on the MICCAI 2015 challenge dataset demonstrate that OrganNet2.5D achieves promising performance compared to state-of-the-art methods.

eess.IV↗

arXiv2021

SDSS-IV MaNGA: the physical origin of off-galaxy H$α$ blobs in the local Universe

H$α$ blobs are off-galaxy emission-line regions with weak or no optical counterparts. They are mostly visible in H$α$ line, appearing as concentrated blobs. Such unusual objects have been rarely observed and studied, and their physical origin is still unclear. We have identified 13 H$α$ blobs in the public data of MaNGA survey, by visually inspecting both the optical images and the spatially resolved maps of H$α$ line for $\sim 4600$ galaxy systems. Among the 13 H$α$ blobs, 2 were reported in previously MaNGA-based studies and 11 are newly discovered. This sample, though still small in size, is by far the largest sample with both deep imaging and integral field spectroscopy. Therefore, for the first time we are able to perform statistical studies to investigate the physical origin of H$α$ blobs. We examine the physical properties of these H$α$ blobs and their associated galaxies, including their morphology, environments, gas-phase metallicity, kinematics of ionized gas, and ionizing sources. We find that the H$α$ blobs in our sample can be broadly divided into two groups. One is associated with interacting/merging galaxy systems, of which the ionization is dominated by shocks or diffuse ionized gas. It is likely that these H$α$ blobs used to be part of their nearby galaxies, but were stripped away at some point due to tidal interactions. The other group is found in gas-rich systems, appearing as low-metallicity star-forming regions that are visually detached from the main galaxy. These H$α$ blobs could be associated with faint disks, spiral arms, or dwarf galaxies.

astro-ph.GA↗

arXiv2021

Annotation-efficient deep learning for automatic medical image segmentation

Automatic medical image segmentation plays a critical role in scientific research and medical care. Existing high-performance deep learning methods typically rely on large training datasets with high-quality manual annotations, which are difficult to obtain in many clinical applications. Here, we introduce Annotation-effIcient Deep lEarning (AIDE), an open-source framework to handle imperfect training datasets. Methodological analyses and empirical evaluations are conducted, and we demonstrate that AIDE surpasses conventional fully-supervised models by presenting better performance on open datasets possessing scarce or noisy annotations. We further test AIDE in a real-life case study for breast tumor segmentation. Three datasets containing 11,852 breast images from three medical centers are employed, and AIDE, utilizing 10% training annotations, consistently produces segmentation maps comparable to those generated by fully-supervised counterparts or provided by independent radiologists. The 10-fold enhanced efficiency in utilizing expert labels has the potential to promote a wide range of biomedical applications.

eess.IV↗

arXiv2021

VIL-100: A New Dataset and A Baseline Model for Video Instance Lane Detection

Lane detection plays a key role in autonomous driving. While car cameras always take streaming videos on the way, current lane detection works mainly focus on individual images (frames) by ignoring dynamics along the video. In this work, we collect a new video instance lane detection (VIL-100) dataset, which contains 100 videos with in total 10,000 frames, acquired from different real traffic scenarios. All the frames in each video are manually annotated to a high-quality instance-level lane annotation, and a set of frame-level and video-level metrics are included for quantitative performance evaluation. Moreover, we propose a new baseline model, named multi-level memory aggregation network (MMA-Net), for video instance lane detection. In our approach, the representation of current frame is enhanced by attentively aggregating both local and global memory features from other frames. Experiments on the new collected dataset show that the proposed MMA-Net outperforms state-of-the-art lane detection methods and video object segmentation methods. We release our dataset and code at https://github.com/yujun0-0/MMA-Net.

cs.CV↗

arXiv2021

MAHGIC: A Model Adapter for the Halo-Galaxy Inter-Connection

We develop a model to establish the interconnection between galaxies and their dark matter halos. We use Principal Component Analysis (PCA) to reduce the dimensionality of both the mass assembly histories of halos/subhalos and the star formation histories of galaxies, and Gradient Boosted Decision Trees (GBDT) to transform halo/subhalo properties into galaxy properties. We use two sets of hydrodynamic simulations to motivate our model architecture and to train the transformation. We then apply the two sets of trained models to dark matter only (DMO) simulations to show that the transformation is reliable and statistically accurate. The model trained by a high-resolution hydrodynamic simulation, or by a set of such simulations implementing the same physics of galaxy formation, can thus be applied to large DMO simulations to make "mock" copies of the hydrodynamic simulation. The model is both flexible and interpretable, which paves the way for future applications in which we will constrain the model using observations at different redshifts simultaneously and explore how galaxies form and evolve in dark matter halos empirically.

astro-ph.GA↗

arXiv2021

A Nonparametric Bayesian Framework for Uncertainty Quantification in Stochastic Simulation

When we use simulation to assess the performance of stochastic systems, the input models used to drive simulation experiments are often estimated from finite real-world data. There exist both input model and simulation estimation uncertainties in the system performance estimates. Without strong prior information on the input models and the system mean response surface, in this paper, we propose a Bayesian nonparametric framework to quantify the impact from both sources of uncertainty. Specifically, since the real-world data often represent the variability caused by various latent sources of uncertainty, Dirichlet Processes Mixtures (DPM) based nonparametric input models are introduced to model a mixture of heterogeneous distributions, which can faithfully capture the important features of real-world data, such as multi-modality and skewness. Bayesian posteriors of flexible input models characterize the input model estimation uncertainty, which automatically accounts for both model selection and parameter value uncertainty. Then, input model estimation uncertainty is propagated to outputs by using direct simulation. Thus, under very general conditions, our framework delivers an empirical credible interval accounting for both input and simulation uncertainties. A variance decomposition is further developed to quantify the relative contributions from both sources of uncertainty. Our approach is supported by rigorous theoretical and empirical study.

stat.ME↗

arXiv2021

Local Manipulation of Polar Skyrmions and Topological Phase Transitions

Topological phases such as polar skyrmions have been a fertile playground for ferroelectric oxide superlattices, with exotic physical phenomena such as negative capacitance. Herein, using phase-field simulations, we demonstrate the local control of the skyrmion phase with electric potential applied through a top electrode. Under a relatively small electric potential, the skyrmions underneath the electrode can be erased and recovered reversibly. A topologically protected transition from the symmetric to asymmetric skyrmion bubbles is observed at the edge of the electrode. While a topological transition to a labyrinthine domain requires a high applied potential, it can switch back to the skyrmion state with a relatively small electric potential. The topological transition from +1 to 0 occurs before the full destruction of the bubble state. It is shown that the shrinking and bursting of the skyrmions leads to a large reduction in the dielectric permittivity, the magnitude of which depends on the size of the electrode.

cond-mat.mtrl-sci↗

arXiv2021

Blind Image Quality Assessment for MRI with A Deep Three-dimensional content-adaptive Hyper-Network

Image Quality Assessment (IQA) is of great value in the workflow of Magnetic Resonance Imaging (MRI)-based analysis. Blind IQA (BIQA) methods are especially required since high-quality reference MRI images are usually not available. Recently, many efforts have been devoted to developing deep learning-based BIQA approaches. However, the performance of these methods is limited due to the utilization of simple content-non-adaptive network parameters and the waste of the important 3D spatial information of the medical images. To address these issues, we design a 3D content-adaptive hyper-network for MRI BIQA. The overall 3D configuration enables the exploration of comprehensive 3D spatial information from MRI images, while the developed content-adaptive hyper-network contributes to the self-adaptive capacity of network parameters and thus, facilitates better BIQA performance. The effectiveness of the proposed method is extensively evaluated on the open dataset, MRIQC. Promising performance is achieved compared with the corresponding baseline and 4 state-of-the-art BIQA methods. We make our code available at \url{https://git.openi.org.cn/SIAT_Wangshanshan/HyS-Net}.

eess.IV↗

arXiv2021

Classification Trees for Imbalanced and Sparse Data: Surface-to-Volume Regularization

Classification algorithms face difficulties when one or more classes have limited training data. We are particularly interested in classification trees, due to their interpretability and flexibility. When data are limited in one or more of the classes, the estimated decision boundaries are often irregularly shaped due to the limited sample size, leading to poor generalization error. We propose a novel approach that penalizes the Surface-to-Volume Ratio (SVR) of the decision set, obtaining a new class of SVR-Tree algorithms. We develop a simple and computationally efficient implementation while proving estimation consistency for SVR-Tree and rate of convergence for an idealized empirical risk minimizer of SVR-Tree. SVR-Tree is compared with multiple algorithms that are designed to deal with imbalance through real data applications.

stat.ME↗

arXiv2021

Estimating dust attenuation from galactic spectra. II. Stellar and gas attenuation in star-forming and diffuse ionized gas regions in MaNGA

We investigate the dust attenuation in both stellar populations and ionized gas in kpc-scale regions in nearby galaxies, using integral field spectroscopy data from MaNGA MPL-9. We identify star-forming (HII) and diffuse ionized gas (DIG) regions from MaNGA datacubes. From the stacked spectrum of each region, we measure the stellar attenuation, $E(B-V)_{\rm star}$, using the technique developed by Li et al.(2020), as well as the gas attenuation, $E(B-V)_{\rm gas}$, from the Balmer decrement. We then examine the correlation of $E(B-V)_{\rm star}$, $E(B-V)_{\rm gas}$, $E(B-V)_{\rm gas}-E(B-V)_{\rm star}$ and $E(B-V)_{\rm star}/E(B-V)_{\rm gas}$ with 16 regional/global properties, and for regions with different $\rm Hα$ surface brightnesses ($Σ_{\rm Hα}$). We find a stronger correlation between $E(B-V)_{\rm star}$ and $E(B-V)_{\rm gas}$ in regions of higher $Σ_{\rm Hα}$. Luminosity-weighted age ($t_L$) is found to be the property that is the most strongly correlated with $E(B-V)_{\rm star}$, and consequently with $E(B-V)_{\rm gas}-E(B-V)_{\rm star}$ and $E(B-V)_{\rm star}/E(B-V)_{\rm gas}$. At fixed $Σ_{\rm Hα}$, $\log_{10}t_L$ is linearly and negatively correlated with $E(B-V)_{\rm star}/E(B-V)_{\rm gas}$ at all ages. Gas-phase metallicity and ionization level are important for the attenuation in the gas. Our results indicate that the ionizing source for DIG regions is likely distributed in the outer-skirt of galaxies, while for HII regions our results can be well explained by the two-component dust model of Charlot & Fall (2000).

astro-ph.GA↗

arXiv2021

Robustness Testing of Language Understanding in Task-Oriented Dialog

Most language understanding models in task-oriented dialog systems are trained on a small amount of annotated training data, and evaluated in a small set from the same distribution. However, these models can lead to system failure or undesirable output when being exposed to natural language perturbation or variation in practice. In this paper, we conduct comprehensive evaluation and analysis with respect to the robustness of natural language understanding models, and introduce three important aspects related to language understanding in real-world dialog systems, namely, language variety, speech characteristics, and noise perturbation. We propose a model-agnostic toolkit LAUG to approximate natural language perturbations for testing the robustness issues in task-oriented dialog. Four data augmentation approaches covering the three aspects are assembled in LAUG, which reveals critical robustness issues in state-of-the-art models. The augmented dataset through LAUG can be used to facilitate future research on the robustness testing of language understanding in task-oriented dialog.

cs.CL↗

arXiv2021

Finding proto-clusters to trace galaxy evolution: I. The finder and its performance

We develop a method to identify proto-clusters based on dark matter halos represented by galaxy groups selected from surveys of galaxies at high redshift. We test the performance of this method on halos in N-body simulations, and find it can correctly identify more than 85% of the true proto-clusters with $\gtrsim$ 95% purity and with mass estimates typical within 0.25 dex from their true values. We show how the information provided by the proto-clusters can be used to link galaxies in present-day clusters of galaxies with their high redshift progenitors. Our tests show that the proto-clusters identified by our method can recover reliably the progenitor stellar mass distribution of galaxies, thereby providing an avenue to investigate the formation and evolution of present-day galaxy clusters and their member galaxies.

astro-ph.GA↗

arXiv2021

Interpretable Biomanufacturing Process Risk and Sensitivity Analyses for Quality-by-Design and Stability Control

While biomanufacturing plays a significant role in supporting the economy and ensuring public health, it faces critical challenges, including complexity, high variability, lengthy lead time, and very limited process data, especially for personalized new cell and gene biotherapeutics. Driven by these challenges, we propose an interpretable semantic bioprocess probabilistic knowledge graph and develop a game theory based risk and sensitivity analyses for production process to facilitate quality-by-design and stability control. Specifically, by exploring the causal relationships and interactions of critical process parameters and quality attributes (CPPs/CQAs), we create a Bayesian network based probabilistic knowledge graph characterizing the complex causal interdependencies of all factors. Then, we introduce a Shapley value based sensitivity analysis, which can correctly quantify the variation contribution from each input factor on the outputs (i.e., productivity, product quality). Since the bioprocess model coefficients are learned from limited process observations, we derive the Bayesian posterior distribution to quantify model uncertainty and further develop the Shapley value based sensitivity analysis to evaluate the impact of estimation uncertainty from each set of model coefficients. Therefore, the proposed bioprocess risk and sensitivity analyses can identify the bottlenecks, guide the reliable process specifications and the most "informative" data collection, and improve production stability.

stat.ML↗

arXiv2021

Star formation histories of massive red spiral galaxies in the local universe

We investigate the star formation histories (SFHs) of massive red spiral galaxies with stellar mass $M_\ast>10^{10.5}M_\odot$, and make comparisons with blue spirals and red ellipticals of similar masses. We make use of the integral field spectroscopy from the SDSS-IV/DR15 MaNGA sample, and estimate spatially resolved SFHs and stellar population properties of each galaxy by applying a Bayesian spectral fitting code to the MaNGA spectra. We find that both red spirals and red ellipticals have experienced only one major star formation episode at early times, and the result is independent of the adopted SFH model. On average, more than half of their stellar masses were formed $>$10 Gyrs ago, and more than 90\% were formed $>6$ Gyrs ago. The two types of galaxies show similarly flat profiles in a variety of stellar population parameters: old stellar ages indicated by $D4000$ (the spectral break at around 4000Å), high stellar metallicities, large Mgb/Fe ratios indicating fast formation, and little stellar dust attenuation. In contrast, although blue spirals also formed their central regions $>$10 Gyrs ago, both their central regions and outer disks continuously form stars over a long timescale. Our results imply that, massive red spirals are likely to share some common processes of formation (and possibly quenching) with massive red ellipticals in the sense that both types were formed at $z > 2$ through a fast formation process.Possible mechanisms for the formation and quenching of massive red spirals are discussed.

astro-ph.GA↗