arXiv subjects

Xiang Li

Publications and source records attributed to Xiang Li.

At least 1,243 records · Page 69Linked to original sources

arXiv2020

Looking Beyond Sentence-Level Natural Language Inference for Downstream Tasks

In recent years, the Natural Language Inference (NLI) task has garnered significant attention, with new datasets and models achieving near human-level performance on it. However, the full promise of NLI -- particularly that it learns knowledge that should be generalizable to other downstream NLP tasks -- has not been realized. In this paper, we study this unfulfilled promise from the lens of two downstream tasks: question answering (QA), and text summarization. We conjecture that a key difference between the NLI datasets and these downstream tasks concerns the length of the premise; and that creating new long premise NLI datasets out of existing QA datasets is a promising avenue for training a truly generalizable NLI model. We validate our conjecture by showing competitive results on the task of QA and obtaining the best reported results on the task of Checking Factual Correctness of Summaries.

cs.CL↗

arXiv2020

Unsupervised Partial Point Set Registration via Joint Shape Completion and Registration

We propose a self-supervised method for partial point set registration. While recent proposed learning-based methods have achieved impressive registration performance on the full shape observations, these methods mostly suffer from performance degradation when dealing with partial shapes. To bridge the performance gaps between partial point set registration with full point set registration, we proposed to incorporate a shape completion network to benefit the registration process. To achieve this, we design a latent code for each pair of shapes, which can be regarded as a geometric encoding of the target shape. By doing so, our model does need an explicit feature embedding network to learn the feature encodings. More importantly, both our shape completion network and the point set registration network take the shared latent codes as input, which are optimized along with the parameters of two decoder networks in the training process. Therefore, the point set registration process can thus benefit from the joint optimization process of latent codes, which are enforced to represent the information of full shape instead of partial ones. In the inference stage, we fix the network parameter and optimize the latent codes to get the optimal shape completion and registration results. Our proposed method is pure unsupervised and does not need any ground truth supervision. Experiments on the ModelNet40 dataset demonstrate the effectiveness of our model for partial point set registration.

cs.CV↗

arXiv2020

GW190521 and the GWTC-1 Events: Implication on the Black Hole Mass Function of Coalescing Binary Black Hole Systems

With the black hole mass function (BHMF; assuming an exponential cutoff at a mass of $\sim 40\,M_\odot$) of coalescing binary black hole systems constructed with the events detected in the O1 run of the advanced LIGO/Virgo network, Liang et al.(2017) predicted that the birth of the lightest intermediate mass black holes (LIMBHs; with a final mass of $\gtrsim 100\,M_\odot$) is very likely to be caught by the advanced LIGO/Virgo detectors in their O3 run. The O1 and O2 observation run data, however, strongly favor a cutoff of the BHMF much sharper than the exponential one. In this work we show that a power-law function followed by a sudden drop at $\sim 40\,M_\odot$ by a factor of $\sim $a few tens and then a new power-law component extending to $\geq 100M_\odot$ are consistent with the O1 and O2 observation run data. With this new BHMF, quite a few LIMBH events can be detected in the O3 observation run of advanced LIGO/Virgo. The first LIMBH born in GW190521, an event detected in the early stage of the O3 run of advanced LIGO/Virgo network, provides additional motivation for our hypothesis.

astro-ph.HE↗

arXiv2020

Inverse Design of Composite Metal Oxide Optical Materials based on Deep Transfer Learning

Optical materials with special optical properties are widely used in a broad span of technologies, from computer displays to solar energy utilization leading to large dataset accumulated from years of extensive materials synthesis and optical characterization. Previously, machine learning models have been developed to predict the optical absorption spectrum from a materials characterization image or vice versa. Herein we propose TLOpt, a transfer learning based inverse optical materials design algorithm for suggesting material compositions with a desired target light absorption spectrum. Our approach is based on the combination of a deep neural network model and global optimization algorithms including a genetic algorithm and Bayesian optimization. A transfer learning strategy is employed to solve the small dataset issue in training the neural network predictor of optical absorption spectrum using the Magpie materials composition descriptor. Our extensive experiments show that our algorithm can inverse design the materials composition with stoichiometry with high accuracy.

cond-mat.mtrl-sci↗

arXiv2020

Object Detection in the Context of Mobile Augmented Reality

In the past few years, numerous Deep Neural Network (DNN) models and frameworks have been developed to tackle the problem of real-time object detection from RGB images. Ordinary object detection approaches process information from the images only, and they are oblivious to the camera pose with regard to the environment and the scale of the environment. On the other hand, mobile Augmented Reality (AR) frameworks can continuously track a camera's pose within the scene and can estimate the correct scale of the environment by using Visual-Inertial Odometry (VIO). In this paper, we propose a novel approach that combines the geometric information from VIO with semantic information from object detectors to improve the performance of object detection on mobile devices. Our approach includes three components: (1) an image orientation correction method, (2) a scale-based filtering approach, and (3) an online semantic map. Each component takes advantage of the different characteristics of the VIO-based AR framework. We implemented the AR-enhanced features using ARCore and the SSD Mobilenet model on Android phones. To validate our approach, we manually labeled objects in image sequences taken from 12 room-scale AR sessions. The results show that our approach can improve on the accuracy of generic object detectors by 12% on our dataset.

cs.CV↗

arXiv2020

Federated Doubly Stochastic Kernel Learning for Vertically Partitioned Data

In a lot of real-world data mining and machine learning applications, data are provided by multiple providers and each maintains private records of different feature sets about common entities. It is challenging to train these vertically partitioned data effectively and efficiently while keeping data privacy for traditional data mining and machine learning algorithms. In this paper, we focus on nonlinear learning with kernels, and propose a federated doubly stochastic kernel learning (FDSKL) algorithm for vertically partitioned data. Specifically, we use random features to approximate the kernel mapping function and use doubly stochastic gradients to update the solutions, which are all computed federatedly without the disclosure of data. Importantly, we prove that FDSKL has a sublinear convergence rate, and can guarantee the data security under the semi-honest assumption. Extensive experimental results on a variety of benchmark datasets show that FDSKL is significantly faster than state-of-the-art federated learning methods when dealing with kernels, while retaining the similar generalization performance.

cs.LG↗

arXiv2020

Robust Image Matching By Dynamic Feature Selection

Estimating dense correspondences between images is a long-standing image under-standing task. Recent works introduce convolutional neural networks (CNNs) to extract high-level feature maps and find correspondences through feature matching. However,high-level feature maps are in low spatial resolution and therefore insufficient to provide accurate and fine-grained features to distinguish intra-class variations for correspondence matching. To address this problem, we generate robust features by dynamically selecting features at different scales. To resolve two critical issues in feature selection,i.e.,how many and which scales of features to be selected, we frame the feature selection process as a sequential Markov decision-making process (MDP) and introduce an optimal selection strategy using reinforcement learning (RL). We define an RL environment for image matching in which each individual action either requires new features or terminates the selection episode by referring a matching score. Deep neural networks are incorporated into our method and trained for decision making. Experimental results show that our method achieves comparable/superior performance with state-of-the-art methods on three benchmarks, demonstrating the effectiveness of our feature selection strategy.

cs.CV↗

arXiv2020

Supervised Community Detection with Line Graph Neural Networks

Traditionally, community detection in graphs can be solved using spectral methods or posterior inference under probabilistic graphical models. Focusing on random graph families such as the stochastic block model, recent research has unified both approaches and identified both statistical and computational detection thresholds in terms of the signal-to-noise ratio. By recasting community detection as a node-wise classification problem on graphs, we can also study it from a learning perspective. We present a novel family of Graph Neural Networks (GNNs) for solving community detection problems in a supervised learning setting. We show that, in a data-driven manner and without access to the underlying generative models, they can match or even surpass the performance of the belief propagation algorithm on binary and multi-class stochastic block models, which is believed to reach the computational threshold. In particular, we propose to augment GNNs with the non-backtracking operator defined on the line graph of edge adjacencies. Our models also achieve good performance on real-world datasets. In addition, we perform the first analysis of the optimization landscape of training linear GNNs for community detection problems, demonstrating that under certain simplifications and assumptions, the loss values at local and global minima are not far apart.

stat.ML↗

arXiv2020

Spectral-change enhancement with prior SNR for the hearing impaired

A previous signal processing algorithm that aimed to enhance spectral changes (SCE) over time showed benefit for hearing-impaired (HI) listeners to recognize speech in background noise. In this work, the previous SCE was manipulated to perform on target-dominant segments, rather than treating all frames equally. Instantaneous signal-to-noise ratios (SNRs) were calculated to determine whether the segments should be processed. Initially, the ideal SNR calculated by the knowledge of premixed signals was introduced to the previous SCE algorithm (SCE-iSNR). Speech intelligibility (SI) and clarity preference were measured for 12 HI listeners in steady speech-spectrum noise (SSN) and six-talk speech (STS) maskers, respectively. The results showed the SCE-iSNR algorithm improved SI significantly for both maskers at high signal-to-masker ratios (SMRs) and for STS masker at low SMRs, while processing effect on speech quality was small. Secondly, the estimated SNR obtained from real mixtures was used, resulting in another SCE-eSNR. SI and subjective rating on naturalness and speech quality were tested for 7 HI subjects. The SCE-eSNR algorithm showed improved SI for SSN masker at high SMRs and for STS masker at low SMRs, as well as better naturalness and speech quality for STS masker. The limitations of applying the algorithms are discussed.

eess.AS↗

arXiv2020

GP-Aligner: Unsupervised Non-rigid Groupwise Point Set Registration Based On Optimized Group Latent Descriptor

In this paper, we propose a novel method named GP-Aligner to deal with the problem of non-rigid groupwise point set registration. Compared to previous non-learning approaches, our proposed method gains competitive advantages by leveraging the power of deep neural networks to effectively and efficiently learn to align a large number of highly deformed 3D shapes with superior performance. Unlike most learning-based methods that use an explicit feature encoding network to extract the per-shape features and their correlations, our model leverages a model-free learnable latent descriptor to characterize the group relationship. More specifically, for a given group we first define an optimizable Group Latent Descriptor (GLD) to characterize the gruopwise relationship among a group of point sets. Each GLD is randomly initialized from a Gaussian distribution and then concatenated with the coordinates of each point of the associated point sets in the group. A neural network-based decoder is further constructed to predict the coherent drifts as the desired transformation from input groups of shapes to aligned groups of shapes. During the optimization process, GP-Aligner jointly updates all GLDs and weight parameters of the decoder network towards the minimization of an unsupervised groupwise alignment loss. After optimization, for each group our model coherently drives each point set towards a middle, common position (shape) without specifying one as the target. GP-Aligner does not require large-scale training data for network training and it can directly align groups of point sets in a one-stage optimization process. GP-Aligner shows both accuracy and computational efficiency improvement in comparison with state-of-the-art methods for groupwise point set registration. Moreover, GP-Aligner is shown great efficiency in aligning a large number of groups of real-world 3D shapes.

cs.CV↗

arXiv2020

Antiferromagnetism of Double Molybdate LiFe(MoO$_4$)$_2$

The magnetic properties of the spin-5/2 double molybdate LiFe(MoO$_4$)$_2$ have been characterized by heat capacity, magnetic susceptibility, and neutron powder diffraction techniques. Unlike the multiferroic system LiFe(MoO$_4$)$_2$ which exhibits two successive magnetic transitions, LiFe(MoO$_4$)$_2$ undergoes only one antiferromagnetic transition at $T_N$ ~ 23.8 K. Its antiferromagnetic magnetic structure with the commensurate propagation vector k = (0, 0.5, 0) has been determined. Density functional theory calculations confirm the antiferromagnetic ground state and provide a numerical estimate of the relevant exchange coupling constants.

cond-mat.mtrl-sci↗

arXiv2020

Simulating multi-exit evacuation using deep reinforcement learning

Conventional simulations on multi-exit indoor evacuation focus primarily on how to determine a reasonable exit based on numerous factors in a changing environment. Results commonly include some congested and other under-utilized exits, especially with massive pedestrians. We propose a multi-exit evacuation simulation based on Deep Reinforcement Learning (DRL), referred to as the MultiExit-DRL, which involves in a Deep Neural Network (DNN) framework to facilitate state-to-action mapping. The DNN framework applies Rainbow Deep Q-Network (DQN), a DRL algorithm that integrates several advanced DQN methods, to improve data utilization and algorithm stability, and further divides the action space into eight isometric directions for possible pedestrian choices. We compare MultiExit-DRL with two conventional multi-exit evacuation simulation models in three separate scenarios: 1) varying pedestrian distribution ratios, 2) varying exit width ratios, and 3) varying open schedules for an exit. The results show that MultiExit-DRL presents great learning efficiency while reducing the total number of evacuation frames in all designed experiments. In addition, the integration of DRL allows pedestrians to explore other potential exits and helps determine optimal directions, leading to the high efficiency of exit utilization.

cs.LG↗

arXiv2020

Observation of the Topologically Originated Edge States in large-gap Quasi-One-Dimensional a-Bi$_4$Br$_4$

Two-dimensional topological insulator features time-reversal-invariant spin-momentum-locked one-dimensional (1D) edge states with a linear energy dispersion. However, experimental access to 1D edge states is still of great challenge and only limited to few techniques to date. Here, by using infrared absorption spectroscopy, we observed robust topologically originated edge states in a-Bi4Br4 belts with definitive signature of strong infrared absorption at belt sides and distinct anisotropy with respect to light polarizations, which is further supported by first-principles calculations. Our work demonstrates for the first time that the infrared spectroscopy can offer a power-efficient approach in experimentally probing 1D edge states of topological materials.

cond-mat.mtrl-sci↗

arXiv2020

Ultralong carrier lifetime of topological edge states in a-Bi4Br4

The rising of quantum spin Hall insulators (QSHI) in two-dimensional (2D) systems has been attracting significant interest in current research, for which the 1D helical edge states, a hallmark of QSHI, are widely expected to be a promising platform for next-generation optoelectronics. However, the dynamics of the 1D edge states has not yet been experimentally addressed. Here, we report the observation of optical response of the topological helical edge states in a-Bi4Br4, using the infrared-pump infrared-probe microscopic spectroscopy. Remarkably, we observe that the carrier lifetime of the helical edge states reaches nanosecond-scale at room temperature, which is about 2 - 3 orders longer than that of most 2D topological surface states and is even comparable with that of the well developed optoelectronics semiconductors used in modern industry. The ultralong carrier lifetime of the topological edge states may be attributed to their helical and 1D nature. Our findings not only provide an ideal material for further investigations of the carrier dynamics of 1D helical edge states but also pave the way for its application in optoelectronics.

cond-mat.mtrl-sci↗

arXiv2020

On the Convergence of FedAvg on Non-IID Data

Federated learning enables a large amount of edge computing devices to jointly learn a model without data sharing. As a leading algorithm in this setting, Federated Averaging (\texttt{FedAvg}) runs Stochastic Gradient Descent (SGD) in parallel on a small subset of the total devices and averages the sequences only once in a while. Despite its simplicity, it lacks theoretical guarantees under realistic settings. In this paper, we analyze the convergence of \texttt{FedAvg} on non-iid data and establish a convergence rate of $\mathcal{O}(\frac{1}{T})$ for strongly convex and smooth problems, where $T$ is the number of SGDs. Importantly, our bound demonstrates a trade-off between communication-efficiency and convergence rate. As user devices may be disconnected from the server, we relax the assumption of full device participation to partial device participation and study different averaging schemes; low device participation rate can be achieved without severely slowing down the learning. Our results indicate that heterogeneity of data slows down the convergence, which matches empirical observations. Furthermore, we provide a necessary condition for \texttt{FedAvg} on non-iid data: the learning rate $η$ must decay, even if full-gradient is used; otherwise, the solution will be $Ω(η)$ away from the optimal.

stat.ML↗

arXiv2020

DeepTracking-Net: 3D Tracking with Unsupervised Learning of Continuous Flow

This paper deals with the problem of 3D tracking, i.e., to find dense correspondences in a sequence of time-varying 3D shapes. Despite deep learning approaches have achieved promising performance for pairwise dense 3D shapes matching, it is a great challenge to generalize those approaches for the tracking of 3D time-varying geometries. In this paper, we aim at handling the problem of 3D tracking, which provides the tracking of the consecutive frames of 3D shapes. We propose a novel unsupervised 3D shape registration framework named DeepTracking-Net, which uses the deep neural networks (DNNs) as auxiliary functions to produce spatially and temporally continuous displacement fields for 3D tracking of objects in a temporal order. Our key novelty is that we present a novel temporal-aware correspondence descriptor (TCD) that captures spatio-temporal essence from consecutive 3D point cloud frames. Specifically, our DeepTracking-Net starts with optimizing a randomly initialized latent TCD. The TCD is then decoded to regress a continuous flow (i.e. a displacement vector field) which assigns a motion vector to every point of time-varying 3D shapes. Our DeepTracking-Net jointly optimizes TCDs and DNNs' weights towards the minimization of an unsupervised alignment loss. Experiments on both simulated and real data sets demonstrate that our unsupervised DeepTracking-Net outperforms the current supervised state-of-the-art method. In addition, we prepare a new synthetic 3D data, named SynMotions, to the 3D tracking and recognition community.

cs.CV↗

arXiv2020

3DMotion-Net: Learning Continuous Flow Function for 3D Motion Prediction

In this paper, we deal with the problem to predict the future 3D motions of 3D object scans from previous two consecutive frames. Previous methods mostly focus on sparse motion prediction in the form of skeletons. While in this paper we focus on predicting dense 3D motions in the from of 3D point clouds. To approach this problem, we propose a self-supervised approach that leverages the power of the deep neural network to learn a continuous flow function of 3D point clouds that can predict temporally consistent future motions and naturally bring out the correspondences among consecutive point clouds at the same time. More specifically, in our approach, to eliminate the unsolved and challenging process of defining a discrete point convolution on 3D point cloud sequences to encode spatial and temporal information, we introduce a learnable latent code to represent the temporal-aware shape descriptor which is optimized during model training. Moreover, a temporally consistent motion Morpher is proposed to learn a continuous flow field which deforms a 3D scan from the current frame to the next frame. We perform extensive experiments on D-FAUST, SCAPE and TOSCA benchmark data sets and the results demonstrate that our approach is capable of handling temporally inconsistent input and produces consistent future 3D motion while requiring no ground truth supervision.

cs.CV↗

arXiv2020

Unsupervised Learning of Global Registration of Temporal Sequence of Point Clouds

Global registration of point clouds aims to find an optimal alignment of a sequence of 2D or 3D point sets. In this paper, we present a novel method that takes advantage of current deep learning techniques for unsupervised learning of global registration from a temporal sequence of point clouds. Our key novelty is that we introduce a deep Spatio-Temporal REPresentation (STREP) feature, which describes the geometric essence of both temporal and spatial relationship of the sequence of point clouds acquired with sensors in an unknown environment. In contrast to the previous practice that treats each time step (pair-wise registration) individually, our unsupervised model starts with optimizing a sequence of latent STREP feature, which is then decoded to a temporally and spatially continuous sequence of geometric transformations to globally align multiple point clouds. We have evaluated our proposed approach over both simulated 2D and real 3D datasets and the experimental results demonstrate that our method can beat other techniques by taking into account the temporal information in deep feature learning.

cs.CV↗