arXiv subjects

Xiao Liu

Publications and source records attributed to Xiao Liu.

At least 505 records · Page 28Linked to original sources

arXiv2019

DF-SLAM: A Deep-Learning Enhanced Visual SLAM System based on Deep Local Features

As the foundation of driverless vehicle and intelligent robots, Simultaneous Localization and Mapping(SLAM) has attracted much attention these days. However, non-geometric modules of traditional SLAM algorithms are limited by data association tasks and have become a bottleneck preventing the development of SLAM. To deal with such problems, many researchers seek to Deep Learning for help. But most of these studies are limited to virtual datasets or specific environments, and even sacrifice efficiency for accuracy. Thus, they are not practical enough. We propose DF-SLAM system that uses deep local feature descriptors obtained by the neural network as a substitute for traditional hand-made features. Experimental results demonstrate its improvements in efficiency and stability. DF-SLAM outperforms popular traditional SLAM systems in various scenes, including challenging scenes with intense illumination changes. Its versatility and mobility fit well into the need for exploring new environments. Since we adopt a shallow network to extract local descriptors and remain others the same as original SLAM systems, our DF-SLAM can still run in real-time on GPU.

cs.CV↗

arXiv2019

Read, Watch, and Move: Reinforcement Learning for Temporally Grounding Natural Language Descriptions in Videos

The task of video grounding, which temporally localizes a natural language description in a video, plays an important role in understanding videos. Existing studies have adopted strategies of sliding window over the entire video or exhaustively ranking all possible clip-sentence pairs in a pre-segmented video, which inevitably suffer from exhaustively enumerated candidates. To alleviate this problem, we formulate this task as a problem of sequential decision making by learning an agent which regulates the temporal grounding boundaries progressively based on its policy. Specifically, we propose a reinforcement learning based framework improved by multi-task learning and it shows steady performance gains by considering additional supervised boundary information during training. Our proposed framework achieves state-of-the-art performance on ActivityNet'18 DenseCaption dataset and Charades-STA dataset while observing only 10 or less clips per video.

cs.CV↗

arXiv2019

CerfGAN: A Compact, Effective, Robust, and Fast Model for Unsupervised Multi-Domain Image-to-Image Translation

In this paper, we aim at solving the multi-domain image-to-image translation problem with a unified model in an unsupervised manner. The most successful work in this area refers to StarGAN, which works well in tasks like face attribute modulation. However, StarGAN is unable to match multiple translation mappings when encountering general translations with very diverse domain shifts. On the other hand, StarGAN adopts an Encoder-Decoder-Discriminator (EDD) architecture, where the model is time-consuming and unstable to train. To this end, we propose a Compact, effective, robust, and fast GAN model, termed CerfGAN, to solve the above problem. In principle, CerfGAN contains a novel component, i.e., a multi-class discriminator (MCD), which gives the model an extremely powerful ability to match multiple translation mappings. To stabilize the training process, MCD also plays a role of the encoder in CerfGAN, which saves a lot of computation and memory costs. We perform extensive experiments to verify the effectiveness of the proposed method. Quantitatively, CerfGAN is demonstrated to handle a serial of image-to-image translation tasks including style transfer, season transfer, face hallucination, etc, where the input images are sampled from diverse domains. The comparisons to several recently proposed approaches demonstrate the superiority and novelty of the proposed method.

cs.CV↗

arXiv2018

Distant Supervision for Relation Extraction with Linear Attenuation Simulation and Non-IID Relevance Embedding

Distant supervision for relation extraction is an efficient method to reduce labor costs and has been widely used to seek novel relational facts in large corpora, which can be identified as a multi-instance multi-label problem. However, existing distant supervision methods suffer from selecting important words in the sentence and extracting valid sentences in the bag. Towards this end, we propose a novel approach to address these problems in this paper. Firstly, we propose a linear attenuation simulation to reflect the importance of words in the sentence with respect to the distances between entities and words. Secondly, we propose a non-independent and identically distributed (non-IID) relevance embedding to capture the relevance of sentences in the bag. Our method can not only capture complex information of words about hidden relations, but also express the mutual information of instances in the bag. Extensive experiments on a benchmark dataset have well-validated the effectiveness of the proposed method.

cs.CL↗

arXiv2018

Development of hp-inverse model by using generalized polynomial chaos

We present a hp-inverse model to estimate a smooth, non-negative source function from a limited number of observations for a two-dimensional linear source inversion problem. A standard least-square inverse model is formulated by using a set of Gaussian radial basis functions (GRBF) on a rectangular mesh system with a uniform grid space. Here, the choice of the mesh system is modeled as a random variable and the generalized polynomial chaos (gPC) expansion is used to represent the random mesh system. It is shown that the convolution of gPC and GRBF provides hierarchical basis functions for the linear source inverse model with the $hp$-refinement capability. We propose a mixed l_1 and l_2 regularization to exploit the hierarchical nature of the basis functions to find a sparse solution. The $hp$-inverse model has an advantage over the standard least-square inverse model when the number of data is limited. It is shown that the hp-inverse model provides a good estimate of the source function even when the number of unknown parameters ($m$) is much larger the number of data ($n$), e.g., m/n > 40.

physics.ao-ph↗

arXiv2018

StNet: Local and Global Spatial-Temporal Modeling for Action Recognition

Despite the success of deep learning for static image understanding, it remains unclear what are the most effective network architectures for the spatial-temporal modeling in videos. In this paper, in contrast to the existing CNN+RNN or pure 3D convolution based approaches, we explore a novel spatial temporal network (StNet) architecture for both local and global spatial-temporal modeling in videos. Particularly, StNet stacks N successive video frames into a \emph{super-image} which has 3N channels and applies 2D convolution on super-images to capture local spatial-temporal relationship. To model global spatial-temporal relationship, we apply temporal convolution on the local spatial-temporal feature maps. Specifically, a novel temporal Xception block is proposed in StNet. It employs a separate channel-wise and temporal-wise convolution over the feature sequence of video. Extensive experiments on the Kinetics dataset demonstrate that our framework outperforms several state-of-the-art approaches in action recognition and can strike a satisfying trade-off between recognition accuracy and model complexity. We further demonstrate the generalization performance of the leaned video representations on the UCF101 dataset.

cs.CV↗

arXiv2018

Solving the 3D High-Frequency Helmholtz Equation using Contour Integration and Polynomial Preconditioning

We propose an iterative solution method for the 3D high-frequency Helmholtz equation that exploits a contour integral formulation of spectral projectors. In this framework, the solution in certain invariant subspaces is approximated by solving complex-shifted linear systems, resulting in faster GMRES iterations due to the restricted spectrum. The shifted systems are solved by exploiting a polynomial fixed-point iteration, which is a robust scheme even if the magnitude of the shift is small. Numerical tests in 3D indicate that $O(n^{1/3})$ matrix-vector products are needed to solve a high-frequency problem with a matrix size $n$ with high accuracy. The method has a small storage requirement, can be applied to both dense and sparse linear systems, and is highly parallelizable.

math.NA↗

arXiv2018

3D Pose Estimation for Fine-Grained Object Categories

Existing object pose estimation datasets are related to generic object types and there is so far no dataset for fine-grained object categories. In this work, we introduce a new large dataset to benchmark pose estimation for fine-grained objects, thanks to the availability of both 2D and 3D fine-grained data recently. Specifically, we augment two popular fine-grained recognition datasets (StanfordCars and CompCars) by finding a fine-grained 3D CAD model for each sub-category and manually annotating each object in images with 3D pose. We show that, with enough training data, a full perspective model with continuous parameters can be estimated using 2D appearance information alone. We achieve this via a framework based on Faster/Mask R-CNN. This goes beyond previous works on category-level pose estimation, which only estimate discrete/continuous viewpoint angles or recover rotation matrices often with the help of key points. Furthermore, with fine-grained 3D models available, we incorporate a dense 3D representation named as location field into the CNN-based pose estimation framework to further improve the performance. The new dataset is available at www.umiacs.umd.edu/~wym/3dpose.html

cs.CV↗

arXiv2018

Fine-grained Video Categorization with Redundancy Reduction Attention

For fine-grained categorization tasks, videos could serve as a better source than static images as videos have a higher chance of containing discriminative patterns. Nevertheless, a video sequence could also contain a lot of redundant and irrelevant frames. How to locate critical information of interest is a challenging task. In this paper, we propose a new network structure, known as Redundancy Reduction Attention (RRA), which learns to focus on multiple discriminative patterns by sup- pressing redundant feature channels. Specifically, it firstly summarizes the video by weight-summing all feature vectors in the feature maps of selected frames with a spatio-temporal soft attention, and then predicts which channels to suppress or to enhance according to this summary with a learned non-linear transform. Suppression is achieved by modulating the feature maps and threshing out weak activations. The updated feature maps are then used in the next iteration. Finally, the video is classified based on multiple summaries. The proposed method achieves out- standing performances in multiple video classification datasets. Further- more, we have collected two large-scale video datasets, YouTube-Birds and YouTube-Cars, for future researches on fine-grained video categorization. The datasets are available at http://www.cs.umd.edu/~chenzhu/fgvc.

cs.CV↗

arXiv2018

Jointly Multiple Events Extraction via Attention-based Graph Information Aggregation

Event extraction is of practical utility in natural language processing. In the real world, it is a common phenomenon that multiple events existing in the same sentence, where extracting them are more difficult than extracting a single event. Previous works on modeling the associations between events by sequential modeling methods suffer a lot from the low efficiency in capturing very long-range dependencies. In this paper, we propose a novel Jointly Multiple Events Extraction (JMEE) framework to jointly extract multiple event triggers and arguments by introducing syntactic shortcut arcs to enhance information flow and attention-based graph convolution networks to model graph information. The experiment results demonstrate that our proposed framework achieves competitive results compared with state-of-the-art methods.

cs.CL↗

arXiv2018

Improving Annotation for 3D Pose Dataset of Fine-Grained Object Categories

Existing 3D pose datasets of object categories are limited to generic object types and lack of fine-grained information. In this work, we introduce a new large-scale dataset that consists of 409 fine-grained categories and 31,881 images with accurate 3D pose annotation. Specifically, we augment three existing fine-grained object recognition datasets (StanfordCars, CompCars and FGVC-Aircraft) by finding a specific 3D model for each sub-category from ShapeNet and manually annotating each 2D image by adjusting a full set of 7 continuous perspective parameters. Since the fine-grained shapes allow 3D models to better fit the images, we further improve the annotation quality by initializing from the human annotation and conducting local search of the pose parameters with the objective of maximizing the IoUs between the projected mask and the segmentation reference estimated from state-of-the-art deep Convolutional Neural Networks (CNNs). We provide full statistics of the annotations with qualitative and quantitative comparisons suggesting that our dataset can be a complementary source for studying 3D pose estimation. The dataset can be downloaded at http://users.umiacs.umd.edu/~wym/3dpose.html.

cs.CV↗

arXiv2018

Semi-analytical calculation of gluon fragmentation into ${^{1}\hspace{-0.6mm}S_{0}^{[1,8]}}$ quarkonia at next-to-leading order

We calculate the NLO corrections for the gluon fragmentation functions to a heavy quark-antiquark pair in ${^{1}\hspace{-0.6mm}S_{0}^{[1]}}$ or ${^{1}\hspace{-0.6mm}S_{0}^{[8]}}$ state within NRQCD factorization. We use integration-by-parts reduction to reduce the original expression to simpler master integrals (MIs), and then set up differential equations for these MIs. After calculating the boundary conditions, MIs can be obtained by solving the differential equations numerically. Our results are expressed in terms of asymptotic expansions at singular points of $z$ (light-cone momentum fraction carried by the quark-antiquark pair), which can not only give FFs results with very high precision at any value of $z$, but also provide fully analytical structure at these singularities. We find that the NLO corrections are significant, with K-factors larger than 2 in most regions. The NLO corrections may have important impact on heavy quarkonia (e.g. $η_c$ and $J/ψ$) production at the LHC.

hep-ph↗

arXiv2018

Thermal coupling of silicon oscillators in cryogen-free dilution refrigerators

Silicon double paddle osillators (DPO) have been successfully used for measuring the elastic properties of amorphous films down to 10 mK (see for example \cite{Fefferman16,Liu14}). Until now, our group has used a wet dilution refrigerator for the lowest temperature measurements. We present measurements carried out on a Bluefors cryogen-free dilution refrigerator that demonstrate an extreme sensitivity of the thermal coupling of the DPO to its environment. These measurements show that it is necessary to enclose the DPO in a shield at the mixing chamber (MXC) temperature. Any gaps in the shield limit its effectiveness, even if there is no line-of-sight path to the DPO. In the absence of a cryogenic hermetic shield surrounding the DPO, turning off the pulse tube while maintaining the MXC and still temperatures leads to heating of the DPO. This demonstrates that any heating of the sample due to pulse tube vibrations is a less important effect.

physics.ins-det↗

arXiv2018

Multiplexed storage and real-time manipulation based on a multiple-degree-of-freedom quantum memory

The faithful storage and coherent manipulation of quantum states with matter-systems enable the construction of large-scale quantum networks based on quantum repeater. To achieve useful communication rates, highly multimode quantum memories will be required to construct a multiplexed quantum repeater. Here, we present the first demonstration of the on-demand storage of orbital-angular-momentum states with weak coherent pulses at the single-photon-level in a rare-earth-ion doped crystal. Through the combination of this spatial degree-of-freedom with temporal and spectral degrees of freedom, we create a multiple-degree-of-freedom memory with high multimode capacity. This device can serve as a quantum mode converter with high fidelity, which is the fundamental requirements for the construction of a multiplexed quantum repeater. This device further enables essentially arbitrary spectral and temporal manipulations of spatial-qutrit-encoded photonic pulses in real-time. Therefore, the developed quantum memory can serve as a building block for scalable photonic quantum information processing architectures.

quant-ph↗

arXiv2018

Anomaly Detection of Complex Networks Based on Intuitionistic Fuzzy Set Ensemble

Ensemble learning for anomaly detection of data structured into complex network has been barely studied due to the inconsistent performance of complex network characteristics and lack of inherent objective function. In this paper, we propose the IFSAD, a new two-phase ensemble method for anomaly detection based on intuitionistic fuzzy set, and applies it to the abnormal behavior detection problem in temporal complex networks. First, it constructs the intuitionistic fuzzy set of single network characteristic which quantifies the degree of membership, non-membership and hesitation of each of network characteristic to the defined linguistic variables so that makes the unuseful or noise characteristics become part of the detection. To build an objective intuitionistic fuzzy relationship, we propose an Gaussian distribution-based membership function which gives a variable hesitation degree. Then, for the fuzzification of multiple network characteristics, the intuitionistic fuzzy weighted geometric operator is adopted to fuse multiple IFSs and to avoid the inconsistent of multiple characteristics. Finally, the score function and precision function are used to sort the fused IFS. Finally we carried out extensive experiments on several complex network datasets for anomaly detection, and the results demonstrate the superiority of our method to state-of-the-art approaches, validating the effectiveness of our method.

cs.SI↗

arXiv2018

YouTube for Patient Education: A Deep Learning Approach for Understanding Medical Knowledge from User-Generated Videos

YouTube presents an unprecedented opportunity to explore how machine learning methods can improve healthcare information dissemination. We propose an interdisciplinary lens that synthesizes machine learning methods with healthcare informatics themes to address the critical issue of developing a scalable algorithmic solution to evaluate videos from a health literacy and patient education perspective. We develop a deep learning method to understand the level of medical knowledge encoded in YouTube videos. Preliminary results suggest that we can extract medical knowledge from YouTube videos and classify videos according to the embedded knowledge with satisfying performance. Deep learning methods show great promise in knowledge extraction, natural language understanding, and image classification, especially in an era of patient-centric care and precision medicine.

cs.CV↗

arXiv2018

Intermodal group velocity engineering for broadband nonlinear optics

Interest in the nonlinear properties of multi-mode optical waveguides has seen a recent resurgence on account of the large dimensionality afforded by the platform. However, a perceived fundamental limitation of intermodal parametric interactions - that they are impractically narrowband - has yet to be solved. Here we show that by engineering the relative group velocity within the discrete spatial degree of freedom, we can tailor the phase matching bandwidth of intermodal parametric nonlinearities. We demonstrate group-velocity-tailored four-wave mixing between the $LP_{0,4}$ and $LP_{0,5}$ modes of a multi-mode fiber with unprecedented gain bandwidths (>60 nm at ~1550 nm). As evidence of the technological utility of this methodology, we seed this process to generate a high-peak-power wavelength-tunable fiber laser in the Ti:Sapphire wavelength regime. More generally, with the combination of intermodal interactions, which dramatically expand the phase matching degrees of freedom for nonlinear optics, and intermodal group velocity engineering, which enables tailoring the bandwidth of such interactions, we showcase a platform for nonlinear optics that can be broadband while being wavelength agnostic.

physics.optics↗

arXiv2018

Probing the collective dynamics of nuclear spin bath in a rare-earth ion doped crystal

Probing collective spin dynamics is a current challenge in the field of magnetic resonance spectroscopy and has important applications in material analysis and quantum information protocols. Recently, the rare-earth ion doped crystals are an attractive candidate for making long-lived quantum memory. Further enhancement of its performance would benefit from the direct knowledge on the dynamics of nuclear-spin bath in the crystal. Here we detect the collective dynamics of nuclear-spin bath located around the rare-earth ions in a crystal using dynamical decoupling spectroscopy method. From the measured spectrum, we analyze the configuration of the spin bath and characterize the flip-flop time between two correlated nuclear spins in a long time scale ($\sim $1s). Furthermore, we experimentally demonstrate that the rare-earth ions can serve as a magnetic quantum sensor for external magnetic field. These results suggest that the rare-earth ion is a useful probe for complex spin dynamics in solids and enable quantum sensing in the low-frequency regime, revealing promising possibilities for applications in diverse fields.

quant-ph↗