arXiv subjects

Yan Xu

Publications and source records attributed to Yan Xu.

At least 199 records · Page 11Linked to original sources

arXiv2022

WSSS4LUAD: Grand Challenge on Weakly-supervised Tissue Semantic Segmentation for Lung Adenocarcinoma

Lung cancer is the leading cause of cancer death worldwide, and adenocarcinoma (LUAD) is the most common subtype. Exploiting the potential value of the histopathology images can promote precision medicine in oncology. Tissue segmentation is the basic upstream task of histopathology image analysis. Existing deep learning models have achieved superior segmentation performance but require sufficient pixel-level annotations, which is time-consuming and expensive. To enrich the label resources of LUAD and to alleviate the annotation efforts, we organize this challenge WSSS4LUAD to call for the outstanding weakly-supervised semantic segmentation (WSSS) techniques for histopathology images of LUAD. Participants have to design the algorithm to segment tumor epithelial, tumor-associated stroma and normal tissue with only patch-level labels. This challenge includes 10,091 patch-level annotations (the training set) and over 130 million labeled pixels (the validation and test sets), from 87 WSIs (67 from GDPH, 20 from TCGA). All the labels were generated by a pathologist-in-the-loop pipeline with the help of AI models and checked by the label review board. Among 532 registrations, 28 teams submitted the results in the test phase with over 1,000 submissions. Finally, the first place team achieved mIoU of 0.8413 (tumor: 0.8389, stroma: 0.7931, normal: 0.8919). According to the technical reports of the top-tier teams, CAM is still the most popular approach in WSSS. Cutmix data augmentation has been widely adopted to generate more reliable samples. With the success of this challenge, we believe that WSSS approaches with patch-level annotations can be a complement to the traditional pixel annotations while reducing the annotation efforts. The entire dataset has been released to encourage more researches on computational pathology in LUAD and more novel WSSS techniques.

eess.IV↗

arXiv2022

Can Question Rewriting Help Conversational Question Answering?

Question rewriting (QR) is a subtask of conversational question answering (CQA) aiming to ease the challenges of understanding dependencies among dialogue history by reformulating questions in a self-contained form. Despite seeming plausible, little evidence is available to justify QR as a mitigation method for CQA. To verify the effectiveness of QR in CQA, we investigate a reinforcement learning approach that integrates QR and CQA tasks and does not require corresponding QR datasets for targeted CQA. We find, however, that the RL method is on par with the end-to-end baseline. We provide an analysis of the failure and describe the difficulty of exploiting QR for CQA.

cs.CL↗

arXiv2022

RNNPose: Recurrent 6-DoF Object Pose Refinement with Robust Correspondence Field Estimation and Pose Optimization

6-DoF object pose estimation from a monocular image is challenging, and a post-refinement procedure is generally needed for high-precision estimation. In this paper, we propose a framework based on a recurrent neural network (RNN) for object pose refinement, which is robust to erroneous initial poses and occlusions. During the recurrent iterations, object pose refinement is formulated as a non-linear least squares problem based on the estimated correspondence field (between a rendered image and the observed image). The problem is then solved by a differentiable Levenberg-Marquardt (LM) algorithm enabling end-to-end training. The correspondence field estimation and pose refinement are conducted alternatively in each iteration to recover the object poses. Furthermore, to improve the robustness to occlusion, we introduce a consistency-check mechanism based on the learned descriptors of the 3D model and observed 2D images, which downweights the unreliable correspondences during pose optimization. Extensive experiments on LINEMOD, Occlusion-LINEMOD, and YCB-Video datasets validate the effectiveness of our method and demonstrate state-of-the-art performance.

cs.CV↗

arXiv2022

Robust Self-Supervised LiDAR Odometry via Representative Structure Discovery and 3D Inherent Error Modeling

The correct ego-motion estimation basically relies on the understanding of correspondences between adjacent LiDAR scans. However, given the complex scenarios and the low-resolution LiDAR, finding reliable structures for identifying correspondences can be challenging. In this paper, we delve into structure reliability for accurate self-supervised ego-motion estimation and aim to alleviate the influence of unreliable structures in training, inference and mapping phases. We improve the self-supervised LiDAR odometry substantially from three aspects: 1) A two-stage odometry estimation network is developed, where we obtain the ego-motion by estimating a set of sub-region transformations and averaging them with a motion voting mechanism, to encourage the network focusing on representative structures. 2) The inherent alignment errors, which cannot be eliminated via ego-motion optimization, are down-weighted in losses based on the 3D point covariance estimations. 3) The discovered representative structures and learned point covariances are incorporated in the mapping module to improve the robustness of map construction. Our two-frame odometry outperforms the previous state of the arts by 16%/12% in terms of translational/rotational errors on the KITTI dataset and performs consistently well on the Apollo-Southbay datasets. We can even rival the fully supervised counterparts with our mapping module and more unlabeled training data.

cs.CV↗

arXiv2022

High-throughput discovery of chemical structure-polarity relationships combining automation and machine learning techniques

As an essential attribute of organic compounds, polarity has a profound influence on many molecular properties such as solubility and phase transition temperature. Thin layer chromatography (TLC) represents a commonly used technique for polarity measurement. However, current TLC analysis presents several problems, including the need for a large number of attempts to obtain suitable conditions, as well as irreproducibility due to non-standardization. Herein, we describe an automated experiment system for TLC analysis. This system is designed to conduct TLC analysis automatically, facilitating high-throughput experimentation by collecting large experimental data under standardized conditions. Using these datasets, machine learning (ML) methods are employed to construct surrogate models correlating organic compounds' structures and their polarity using retardation factor (Rf). The trained ML models are able to predict the Rf value curve of organic compounds with high accuracy. Furthermore, the constitutive relationship between the compound and its polarity can also be discovered through these modeling methods, and the underlying mechanism is rationalized through adsorption theories. The trained ML models not only reduce the need for empirical optimization currently required for TLC analysis, but also provide general guidelines for the selection of conditions, making TLC an easily accessible tool for the broad scientific community.

physics.chem-ph↗

arXiv2022

SelfVoxeLO: Self-supervised LiDAR Odometry with Voxel-based Deep Neural Networks

Recent learning-based LiDAR odometry methods have demonstrated their competitiveness. However, most methods still face two substantial challenges: 1) the 2D projection representation of LiDAR data cannot effectively encode 3D structures from the point clouds; 2) the needs for a large amount of labeled data for training limit the application scope of these methods. In this paper, we propose a self-supervised LiDAR odometry method, dubbed SelfVoxeLO, to tackle these two difficulties. Specifically, we propose a 3D convolution network to process the raw LiDAR data directly, which extracts features that better encode the 3D geometric patterns. To suit our network to self-supervised learning, we design several novel loss functions that utilize the inherent properties of LiDAR point clouds. Moreover, an uncertainty-aware mechanism is incorporated in the loss functions to alleviate the interference of moving objects/noises. We evaluate our method's performances on two large-scale datasets, i.e., KITTI and Apollo-SouthBay. Our method outperforms state-of-the-art unsupervised methods by 27%/32% in terms of translational/rotational errors on the KITTI dataset and also performs well on the Apollo-SouthBay dataset. By including more unlabelled training data, our method can further improve performance comparable to the supervised methods.

cs.CV↗

arXiv2022

Numerical simulation of singularity propagation modeled by linear convection equations with spatially heterogeneous nonlocal interactions

We study the propagation of singularities in solutions of linear convection equations with spatially heterogeneous nonlocal interactions. A spatially varying nonlocal horizon parameter is adopted in the model, which measures the range of nonlocal interactions. Via heterogeneous localization, this can lead to the seamless coupling of the local and nonlocal models. We are interested in understanding the impact on singularity propagation due to the heterogeneities of nonlocal horizon and the local and nonlocal transition. We first analytically derive equations to characterize the propagation of different types of singularities for various forms of nonlocal horizon parameters in the nonlocal regime. We then use asymptotically compatible schemes to discretize the equations and carry out numerical simulations to illustrate the propagation patterns in different scenarios.

math.NA↗

arXiv2021

Multi-Passband Observations of A Solar Flare over the He I 10830 Å line

This study presents a C3.0 flare observed by the BBSO/GST and IRIS, on 2018-May-28 around 17:10 UT. The Near Infrared Imaging Spectropolarimeter (NIRIS) of GST was set to spectral imaging mode to scan five spectral positions at $\pm$ 0.8 Å, $\pm$ 0.4 Åand line center of He I 10830. At the flare ribbon's leading edge the line is observed to undergo enhanced absorption, while the rest of the ribbon is observed to be in emission. When in emission, the contrast compared to the pre-flare ranges from about $30~\%$ to nearly $100~\%$ at different spectral positions. Two types of spectra, "convex" shape with higher intensity at line core and "concave" shape with higher emission in the line wings, are found at the trailing and peak flaring areas, respectively. On the ribbon front, negative contrasts, or enhanced absorption, of about $\sim 10\% - 20\%$ appear in all five wavelengths. This observation strongly suggests that the negative flares observed in He I 10830 with mono-filtergram previously were not caused by pure Doppler shifts of this spectral line. Instead, the enhanced absorption appears to be a consequence of flare energy injection, namely non-thermal collisional ionization of helium caused by the precipitation of high energy electrons, as found in our recent numerical modeling results. In addition, though not strictly simultaneous, observations of Mg II from the IRIS spacecraft, show an obvious central reversal pattern at the locations where enhanced absorption of He I 10830 is seen, which is in consistent with previous observations.

astro-ph.SR↗

arXiv2021

Pedestrian Trajectory Prediction via Spatial Interaction Transformer Network

As a core technology of the autonomous driving system, pedestrian trajectory prediction can significantly enhance the function of active vehicle safety and reduce road traffic injuries. In traffic scenes, when encountering with oncoming people, pedestrians may make sudden turns or stop immediately, which often leads to complicated trajectories. To predict such unpredictable trajectories, we can gain insights into the interaction between pedestrians. In this paper, we present a novel generative method named Spatial Interaction Transformer (SIT), which learns the spatio-temporal correlation of pedestrian trajectories through attention mechanisms. Furthermore, we introduce the conditional variational autoencoder (CVAE) framework to model the future latent motion states of pedestrians. In particular, the experiments based on large-scale trafc dataset nuScenes [2] show that SIT has an outstanding performance than state-of-the-art (SOTA) methods. Experimental evaluation on the challenging ETH and UCY datasets conrms the robustness of our proposed model

cs.CV↗

arXiv2021

Whole Brain Segmentation with Full Volume Neural Network

Whole brain segmentation is an important neuroimaging task that segments the whole brain volume into anatomically labeled regions-of-interest. Convolutional neural networks have demonstrated good performance in this task. Existing solutions, usually segment the brain image by classifying the voxels, or labeling the slices or the sub-volumes separately. Their representation learning is based on parts of the whole volume whereas their labeling result is produced by aggregation of partial segmentation. Learning and inference with incomplete information could lead to sub-optimal final segmentation result. To address these issues, we propose to adopt a full volume framework, which feeds the full volume brain image into the segmentation network and directly outputs the segmentation result for the whole brain volume. The framework makes use of complete information in each volume and can be implemented easily. An effective instance in this framework is given subsequently. We adopt the $3$D high-resolution network (HRNet) for learning spatially fine-grained representations and the mixed precision training scheme for memory-efficient training. Extensive experiment results on a publicly available $3$D MRI brain dataset show that our proposed model advances the state-of-the-art methods in terms of segmentation performance. Source code is publicly available at https://github.com/microsoft/VoxHRNet.

eess.IV↗

arXiv2021

Testing and Support Recovery of Correlation Structures for Matrix-Valued Observations with an Application to Stock Market Data

Estimation of the covariance matrix of asset returns is crucial to portfolio construction. As suggested by economic theories, the correlation structure among assets differs between emerging markets and developed countries. It is therefore imperative to make rigorous statistical inference on correlation matrix equality between the two groups of countries. However, if the traditional vector-valued approach is undertaken, such inference is either infeasible due to limited number of countries comparing to the relatively abundant assets, or invalid due to the violations of temporal independence assumption. This highlights the necessity of treating the observations as matrix-valued rather than vector-valued. With matrix-valued observations, our problem of interest can be formulated as statistical inference on covariance structures under sub-Gaussian distributions, i.e., testing non-correlation and correlation equality, as well as the corresponding support estimations. We develop procedures that are asymptotically optimal under some regularity conditions. Simulation results demonstrate the computational and statistical advantages of our procedures over certain existing state-of-the-art methods for both normal and non-normal distributions. Application of our procedures to stock market data reveals interesting patterns and validates several economic propositions via rigorous statistical testing.

stat.ME↗

arXiv2021

Tracing Halpha Fibrils through Bayesian Deep Learning

We present a new deep learning method, dubbed FibrilNet, for tracing chromospheric fibrils in Halpha images of solar observations. Our method consists of a data pre-processing component that prepares training data from a threshold-based tool, a deep learning model implemented as a Bayesian convolutional neural network for probabilistic image segmentation with uncertainty quantification to predict fibrils, and a post-processing component containing a fibril-fitting algorithm to determine fibril orientations. The FibrilNet tool is applied to high-resolution Halpha images from an active region (AR 12665) collected by the 1.6 m Goode Solar Telescope (GST) equipped with high-order adaptive optics at the Big Bear Solar Observatory (BBSO). We quantitatively assess the FibrilNet tool, comparing its image segmentation algorithm and fibril-fitting algorithm with those employed by the threshold-based tool. Our experimental results and major findings are summarized as follows. First, the image segmentation results (i.e., detected fibrils) of the two tools are quite similar, demonstrating the good learning capability of FibrilNet. Second, FibrilNet finds more accurate and smoother fibril orientation angles than the threshold-based tool. Third, FibrilNet is faster than the threshold-based tool and the uncertainty maps produced by FibrilNet not only provide a quantitative way to measure the confidence on each detected fibril, but also help identify fibril structures that are not detected by the threshold-based tool but are inferred through machine learning. Finally, we apply FibrilNet to full-disk Halpha images from other solar observatories and additional high-resolution Halpha images collected by BBSO/GST, demonstrating the tool's usability in diverse datasets.

astro-ph.SR↗

arXiv2021

EDGE COVID-19: A Web Platform to generate submission-ready genomes for SARS-CoV-2 sequencing efforts

Genomics has become an essential technology for surveilling emerging infectious disease outbreaks. A wide range of technologies and strategies for pathogen genome enrichment and sequencing are being used by laboratories worldwide, together with different, and sometimes ad hoc, analytical procedures for generating genome sequences. As a result, public repositories now contain non-standard entries of varying quality. A standardized analytical process for consensus genome sequence determination, particularly for outbreaks such as the ongoing COVID-19 pandemic, is critical to provide a solid genomic basis for epidemiological analyses and well-informed decision making. To address this need, we have developed a bioinformatic workflow to standardize the analysis of SARS-CoV-2 sequencing data generated with either the Illumina or Oxford Nanopore platforms. Using an intuitive web-based interface, this workflow automates SARS-CoV-2 reference-based genome assembly, variant calling, lineage determination, and provides the ability to submit the consensus sequence and necessary metadata to GenBank or GISAID. Given a raw Illumina or Oxford Nanopore FASTQ read file, this web-based platform enables non-bioinformatics experts to automatically produce a SARS-CoV-2 genome that is ready for submission to GISAID or GenBank. Availability:https://edge-covid19.edgebioinformatics.org;https://github.com/LANL-Bioinformatics/EDGE/tree/SARS-CoV2

q-bio.GN↗

arXiv2021

SiamSNN: Siamese Spiking Neural Networks for Energy-Efficient Object Tracking

Recently spiking neural networks (SNNs), the third-generation of neural networks has shown remarkable capabilities of energy-efficient computing, which is a promising alternative for deep neural networks (DNNs) with high energy consumption. SNNs have reached competitive results compared to DNNs in relatively simple tasks and small datasets such as image classification and MNIST/CIFAR, while few studies on more challenging vision tasks on complex datasets. In this paper, we focus on extending deep SNNs to object tracking, a more advanced vision task with embedded applications and energy-saving requirements, and present a spike-based Siamese network called SiamSNN. Specifically, we propose an optimized hybrid similarity estimation method to exploit temporal information in the SNNs, and introduce a novel two-status coding scheme to optimize the temporal distribution of output spike trains for further improvements. SiamSNN is the first deep SNN tracker that achieves short latency and low precision loss on the visual object tracking benchmarks OTB2013/2015, VOT2016/2018, and GOT-10k. Moreover, SiamSNN achieves notably low energy consumption and real-time on Neuromorphic chip TrueNorth.

cs.CV↗

arXiv2021

CAiRE in DialDoc21: Data Augmentation for Information-Seeking Dialogue System

Information-seeking dialogue systems, including knowledge identification and response generation, aim to respond to users with fluent, coherent, and informative responses based on users' needs, which. To tackle this challenge, we utilize data augmentation methods and several training techniques with the pre-trained language models to learn a general pattern of the task and thus achieve promising performance. In DialDoc21 competition, our system achieved 74.95 F1 score and 60.74 Exact Match score in subtask 1, and 37.72 SacreBLEU score in subtask 2. Empirical analysis is provided to explain the effectiveness of our approaches.

cs.CL↗

arXiv2021

VS-Net: Voting with Segmentation for Visual Localization

Visual localization is of great importance in robotics and computer vision. Recently, scene coordinate regression based methods have shown good performance in visual localization in small static scenes. However, it still estimates camera poses from many inferior scene coordinates. To address this problem, we propose a novel visual localization framework that establishes 2D-to-3D correspondences between the query image and the 3D map with a series of learnable scene-specific landmarks. In the landmark generation stage, the 3D surfaces of the target scene are over-segmented into mosaic patches whose centers are regarded as the scene-specific landmarks. To robustly and accurately recover the scene-specific landmarks, we propose the Voting with Segmentation Network (VS-Net) to segment the pixels into different landmark patches with a segmentation branch and estimate the landmark locations within each patch with a landmark location voting branch. Since the number of landmarks in a scene may reach up to 5000, training a segmentation network with such a large number of classes is both computation and memory costly for the commonly used cross-entropy loss. We propose a novel prototype-based triplet loss with hard negative mining, which is able to train semantic segmentation networks with a large number of labels efficiently. Our proposed VS-Net is extensively tested on multiple public benchmarks and can outperform state-of-the-art visual localization methods. Code and models are available at \href{https://github.com/zju3dv/VS-Net}{https://github.com/zju3dv/VS-Net}.

cs.CV↗

arXiv2021

He I 10830 Å Dimming During Solar Flares, I. The Crucial Role of Non-Thermal Collisional Ionisations

While solar flares are predominantly characterised by an intense broadband enhancement to the solar radiative output, certain spectral lines and continua will, in theory, exhibit flare-induced dimmings. Observations of transitions of orthohelium He I $λλ10830$Å and the He I D3 lines have shown evidence of such dimming, usually followed by enhanced emission. It has been suggested that non-thermal collisional ionisation of helium by an electron beam, followed by recombinations to orthohelium, is responsible for overpopulating the those levels, leading to stronger absorption. However it has not been possible observationally to preclude the possibility of overpopulating orthohelium via enhanced photoionisation of He I by EUV irradiance from the flaring corona followed by recombinations. Here we present radiation hydrodynamics simulations of non-thermal electron beam-driven flares where (1) both non-thermal collisional ionisation of Helium and coronal irradiance are included, and (2) only coronal irradiance is included. A grid of simulations covering a range of total energies deposited by the electron beam, and a range of non-thermal electron beam low-energy cutoff values, were simulated. In order to obtain flare-induced dimming of the He I 10830Å line it was necessary for non-thermal collisional ionisations to be present. The effect was more prominent in flares with larger low-energy cutoff values and longer lived in weaker flares and flares with a more gradual energy deposition timescale. These results demonstrate the usefulness of orthohelium line emission as a diagnostic of flare energy transport.

astro-ph.SR↗

arXiv2021

LIFE: Lighting Invariant Flow Estimation

We tackle the problem of estimating flow between two images with large lighting variations. Recent learning-based flow estimation frameworks have shown remarkable performance on image pairs with small displacement and constant illuminations, but cannot work well on cases with large viewpoint change and lighting variations because of the lack of pixel-wise flow annotations for such cases. We observe that via the Structure-from-Motion (SfM) techniques, one can easily estimate relative camera poses between image pairs with large viewpoint change and lighting variations. We propose a novel weakly supervised framework LIFE to train a neural network for estimating accurate lighting-invariant flows between image pairs. Sparse correspondences are conventionally established via feature matching with descriptors encoding local image contents. However, local image contents are inevitably ambiguous and error-prone during the cross-image feature matching process, which hinders downstream tasks. We propose to guide feature matching with the flows predicted by LIFE, which addresses the ambiguous matching by utilizing abundant context information in the image pairs. We show that LIFE outperforms previous flow learning frameworks by large margins in challenging scenarios, consistently improves feature matching, and benefits downstream tasks.

cs.CV↗