arXiv subjects

Dong Liu

Publications and source records attributed to Dong Liu.

At least 271 records · Page 15Linked to original sources

arXiv2021

Deep Learning Aided Packet Routing in Aeronautical Ad-Hoc Networks Relying on Real Flight Data: From Single-Objective to Near-Pareto Multi-Objective Optimization

Data packet routing in aeronautical ad-hoc networks (AANETs) is challenging due to their high-dynamic topology. In this paper, we invoke deep learning (DL) to assist routing in AANETs. We set out from the single objective of minimizing the end-to-end (E2E) delay. Specifically, a deep neural network (DNN) is conceived for mapping the local geographic information observed by the forwarding node into the information required for determining the optimal next hop. The DNN is trained by exploiting the regular mobility pattern of commercial passenger airplanes from historical flight data. After training, the DNN is stored by each airplane for assisting their routing decisions during flight relying solely on local geographic information. Furthermore, we extend the DL-aided routing algorithm to a multi-objective scenario, where we aim for simultaneously minimizing the delay, maximizing the path capacity, and maximizing the path lifetime. Our simulation results based on real flight data show that the proposed DL-aided routing outperforms existing position-based routing protocols in terms of its E2E delay, path capacity as well as path lifetime, and it is capable of approaching the Pareto front that is obtained using global link information.

cs.NI↗

arXiv2021

Deep Reinforcement Learning Aided Packet-Routing For Aeronautical Ad-Hoc Networks Formed by Passenger Planes

Data packet routing in aeronautical ad-hoc networks (AANETs) is challenging due to their high-dynamic topology. In this paper, we invoke deep reinforcement learning for routing in AANETs aiming at minimizing the end-to-end (E2E) delay. Specifically, a deep Q-network (DQN) is conceived for capturing the relationship between the optimal routing decision and the local geographic information observed by the forwarding node. The DQN is trained in an offline manner based on historical flight data and then stored by each airplane for assisting their routing decisions during flight. To boost the learning efficiency and the online adaptability of the proposed DQN-routing, we further exploit the knowledge concerning the system's dynamics by using a deep value network (DVN) conceived with a feedback mechanism. Our simulation results show that both DQN-routing and DVN-routing achieve lower E2E delay than the benchmark protocol, and DVN-routing performs similarly to the optimal routing that relies on perfect global information.

cs.NI↗

arXiv2021

iWave3D: End-to-end Brain Image Compression with Trainable 3-D Wavelet Transform

With the rapid development of whole brain imaging technology, a large number of brain images have been produced, which puts forward a great demand for efficient brain image compression methods. At present, the most commonly used compression methods are all based on 3-D wavelet transform, such as JP3D. However, traditional 3-D wavelet transforms are designed manually with certain assumptions on the signal, but brain images are not as ideal as assumed. What's more, they are not directly optimized for compression task. In order to solve these problems, we propose a trainable 3-D wavelet transform based on the lifting scheme, in which the predict and update steps are replaced by 3-D convolutional neural networks. Then the proposed transform is embedded into an end-to-end compression scheme called iWave3D, which is trained with a large amount of brain images to directly minimize the rate-distortion loss. Experimental results demonstrate that our method outperforms JP3D significantly by 2.012 dB in terms of average BD-PSNR.

eess.IV↗

arXiv2021

Simultaneous Cluster Structure Learning and Estimation of Heterogeneous Graphs for Matrix-variate fMRI Data

Graphical models play an important role in neuroscience studies, particularly in brain connectivity analysis. Typically, observations/samples are from several heterogenous groups and the group membership of each observation/sample is unavailable, which poses a great challenge for graph structure learning. In this article, we propose a method which can achieve Simultaneous Clustering and Estimation of Heterogeneous Graphs (briefly denoted as SCEHG) for matrix-variate function Magnetic Resonance Imaging (fMRI) data. Unlike the conventional clustering methods which rely on the mean differences of various groups, the proposed SCEHG method fully exploits the group differences of conditional dependence relationships among brain regions for learning cluster structure. In essence, by constructing individual-level between-region network measures, we formulate clustering as penalized regression with grouping and sparsity pursuit, which transforms the unsupervised learning into supervised learning. An ADMM algorithm is proposed to solve the corresponding optimization problem. We also propose a generalized criterion to specify the number of clusters. Extensive simulation studies illustrate the superiority of the SCEHG method over some state-of-the-art methods in terms of both clustering and graph recovery accuracy. We also apply the SCEHG procedure to analyze fMRI data associated with ADHD (abbreviated for Attention Deficit Hyperactivity Disorder), which illustrate its empirical usefulness. An R package ``SCEHG" to implement the method is available at https://github.com/heyongstat/SCEHG.

stat.ME↗

arXiv2021

End-to-End Image Compression with Probabilistic Decoding

Lossy image compression is a many-to-one process, thus one bitstream corresponds to multiple possible original images, especially at low bit rates. However, this nature was seldom considered in previous studies on image compression, which usually chose one possible image as reconstruction, e.g. the one with the maximal a posteriori probability. We propose a learned image compression framework to natively support probabilistic decoding. The compressed bitstream is decoded into a series of parameters that instantiate a pre-chosen distribution; then the distribution is used by the decoder to sample and reconstruct images. The decoder may adopt different sampling strategies and produce diverse reconstructions, among which some have higher signal fidelity and some others have better visual quality. The proposed framework is dependent on a revertible neural network-based transform to convert pixels into coefficients that obey the pre-chosen distribution as much as possible. Our code and models will be made publicly available.

eess.IV↗

arXiv2021

Graph-Based 3D Multi-Person Pose Estimation Using Multi-View Images

This paper studies the task of estimating the 3D human poses of multiple persons from multiple calibrated camera views. Following the top-down paradigm, we decompose the task into two stages, i.e. person localization and pose estimation. Both stages are processed in coarse-to-fine manners. And we propose three task-specific graph neural networks for effective message passing. For 3D person localization, we first use Multi-view Matching Graph Module (MMG) to learn the cross-view association and recover coarse human proposals. The Center Refinement Graph Module (CRG) further refines the results via flexible point-based prediction. For 3D pose estimation, the Pose Regression Graph Module (PRG) learns both the multi-view geometry and structural relations between human joints. Our approach achieves state-of-the-art performance on CMU Panoptic and Shelf datasets with significantly lower computation complexity.

cs.CV↗

arXiv2021

Influences of ALD Al$_2$O$_3$ on the surface band-bending of c-plane, Ga-face GaN and the implication to GaN-collector npn heterojunction bipolar transistors

Due to the lack of effective p-type doping in GaN and the adverse effects of surface band-bending of GaN on electron transport, developing practical GaN heterojunction bipolar transistors has been impossible. The recently demonstrated approach of grafting n-type GaN with p-type semiconductors, like Si and GaAs, by employing ultrathin (UO) Al$_2$O$_3$ at the interface of Si/GaN and GaAs/GaN, has shown the feasibility to overcome the poor p-type doping challenge of GaN by providing epitaxy-like interface quality. However, the surface band-bending of GaN that could be influenced by the UO Al2O3 has been unknown. In this work, the band-bending of c-plane, Ga-face GaN with UO Al2O3 deposition at the surface of GaN was studied using X-ray photoelectron spectroscopy (XPS). The study shows that the UO Al2O3 can help in suppressing the upward band-bending of the c-plane, Ga-face GaN with a monotonic reduction trend of the upward band-bending energy from 0.48 eV down to 0.12 eV as the number of UO Al2O3 deposition cycles is increased from 0 to 20 cycles. The study further shows that the band-bending can be mostly recovered after removing the Al2O3 layer, concurring that the change in the density of fixed charge at the GaN surface caused by UO Al2O3 is the main reason for the surface band-bending modulation. The potential implication of the surface band-bending results of AlGaAs/GaAs/GaN npn heterojunction bipolar transistor (HBT) was preliminarily studied via Silvaco(R) simulations.

physics.app-ph↗

arXiv2021

Is Heuristic Sampling Necessary in Training Deep Object Detectors?

To train accurate deep object detectors under the extreme foreground-background imbalance, heuristic sampling methods are always necessary, which either re-sample a subset of all training samples (hard sampling methods, \eg biased sampling, OHEM), or use all training samples but re-weight them discriminatively (soft sampling methods, \eg Focal Loss, GHM). In this paper, we challenge the necessity of such hard/soft sampling methods for training accurate deep object detectors. While previous studies have shown that training detectors without heuristic sampling methods would significantly degrade accuracy, we reveal that this degradation comes from an unreasonable classification gradient magnitude caused by the imbalance, rather than a lack of re-sampling/re-weighting. Motivated by our discovery, we propose a simple yet effective \emph{Sampling-Free} mechanism to achieve a reasonable classification gradient magnitude by initialization and loss scaling. Unlike heuristic sampling methods with multiple hyperparameters, our Sampling-Free mechanism is fully data diagnostic, without laborious hyperparameters searching. We verify the effectiveness of our method in training anchor-based and anchor-free object detectors, where our method always achieves higher detection accuracy than heuristic sampling methods on COCO and PASCAL VOC datasets. Our Sampling-Free mechanism provides a new perspective to address the foreground-background imbalance. Our code is released at \url{https://github.com/ChenJoya/sampling-free}.

cs.CV↗

arXiv2021

Engineered Raman Lasing in Photonic Integrated Chalcogenide Microresonators

Chalcogenide glass (ChG) is an attractive material for integrated nonlinear photonics due to its wide transparency and high nonlinearity, and its capability of being directly deposited and patterned on Silicon wafer substrates. It has a singular Raman effect among amorphous materials. Yet, the Raman lasing performance in high quality and chip integrated ChG microresonators remains unexplored. Here, we demonstrate an engineered Raman lasing dynamic based on home developed photonic integrated high-Q ChG microresonators. With a quality factor above 10^6, we achieve the record-low lasing threshold 3.25 mW among integrated planar photonic platforms. Both the single-mode Raman lasers and a broadband Raman-Kerr comb are observed and characterized, which is dependent on the dispersion of our flexible photonic platform and engineered via tuning the waveguide geometric size. The tunability of such a chipscale Raman laser is also demonstrated through tuning the pump wavelength and tuning the operating temperature on the chip. This allows for the access of single-mode lasing at arbitrary wavelengths in the range 1615-1755 nm. Our results may contribute to the understanding of rich Raman and Kerr nonlinear interactions in dissipative and nonlinear microresonators, and on application aspect, may pave a way to chip-scale efficient Raman lasers that is highly desired in spectroscopic applications in the infrared.

physics.optics↗

arXiv2021

Normalizing Flow based Hidden Markov Models for Classification of Speech Phones with Explainability

In pursuit of explainability, we develop generative models for sequential data. The proposed models provide state-of-the-art classification results and robust performance for speech phone classification. We combine modern neural networks (normalizing flows) and traditional generative models (hidden Markov models - HMMs). Normalizing flow-based mixture models (NMMs) are used to model the conditional probability distribution given the hidden state in the HMMs. Model parameters are learned through judicious combinations of time-tested Bayesian learning methods and contemporary neural network learning methods. We mainly combine expectation-maximization (EM) and mini-batch gradient descent. The proposed generative models can compute likelihood of a data and hence directly suitable for maximum-likelihood (ML) classification approach. Due to structural flexibility of HMMs, we can use different normalizing flow models. This leads to different types of HMMs providing diversity in data modeling capacity. The diversity provides an opportunity for easy decision fusion from different models. For a standard speech phone classification setup involving 39 phones (classes) and the TIMIT dataset, we show that the use of standard features called mel-frequency-cepstral-coeffcients (MFCCs), the proposed generative models, and the decision fusion together can achieve $86.6\%$ accuracy by generative training only. This result is close to state-of-the-art results, for examples, $86.2\%$ accuracy of PyTorch-Kaldi toolkit [1], and $85.1\%$ accuracy using light gated recurrent units [2]. We do not use any discriminative learning approach and related sophisticated features in this article.

cs.LG↗

arXiv2021

Joint Learning of Multiple Differential Networks with fMRI data for Brain Connectivity Alteration Detection

In this study we focus on the problem of joint learning of multiple differential networks with function Magnetic Resonance Imaging (fMRI) data sets from multiple research centers. As the research centers may use different scanners and imaging parameters, joint learning of differential networks with fMRI data from different centers may reflect the underlying mechanism of neurological diseases from different perspectives while capturing the common structures. We transform the task as a penalized logistic regression problem, and exploit sparse group Minimax Concave Penalty (gMCP) to induce common structures among multiple differential networks and the sparse structures of each differential network. To further enhance the empirical performance, we develop an ensemble-learning procedure. We conduct thorough simulation study to assess the finite-sample performance of the proposed method and compare with state-of-the-art alternatives. We apply the proposed method to analyze fMRI datasets related with Attention Deficit Hyperactivity Disorder from various research centers. The identified common hub nodes and differential interaction patterns coincides with the existing experimental studies.

stat.ME↗

arXiv2021

Adaptive Domain-Specific Normalization for Generalizable Person Re-Identification

Although existing person re-identification (Re-ID) methods have shown impressive accuracy, most of them usually suffer from poor generalization on unseen target domain. Thus, generalizable person Re-ID has recently drawn increasing attention, which trains a model on source domains that generalizes well on unseen target domain without model updating. In this work, we propose a novel adaptive domain-specific normalization approach (AdsNorm) for generalizable person Re-ID. It describes unseen target domain as a combination of the known source ones, and explicitly learns domain-specific representation with target distribution to improve the model's generalization by a meta-learning pipeline. Specifically, AdsNorm utilizes batch normalization layers to collect individual source domains' characteristics, and maps source domains into a shared latent space by using these characteristics, where the domain relevance is measured by a distance function of different domain-specific normalization statistics and features. At the testing stage, AdsNorm projects images from unseen target domain into the same latent space, and adaptively integrates the domain-specific features carrying the source distributions by domain relevance for learning more generalizable aggregated representation on unseen target domain. Considering that target domain is unavailable during training, a meta-learning algorithm combined with a customized relation loss is proposed to optimize an effective and efficient ensemble model. Extensive experiments demonstrate that AdsNorm outperforms the state-of-the-art methods. The code is available at: https://github.com/hzphzp/AdsNorm.

cs.CV↗

arXiv2021

Component Edge Connectivity of Hypercube-like Networks

As a generalization of the traditional connectivity, the g-component edge connectivity cλg(G) of a non-complete graph G is the minimum number of edges to be deleted from the graph G such that the resulting graph has at least g components. Hypercube-like networks (HL-networks for short) are obtained by manipulating some pairs of edges in hypercubes, which contain several famous interconnection networks such as twisted cubes, Mobius cubes, crossed cubes, locally twisted cubes. In this paper, we determine the (g + 1)-component edge connectivity of the n-dimensional HL-networks.

math.CO↗

arXiv2021

Simultaneous Navigation and Construction Benchmarking Environments

We need intelligent robots for mobile construction, the process of navigating in an environment and modifying its structure according to a geometric design. In this task, a major robot vision and learning challenge is how to exactly achieve the design without GPS, due to the difficulty caused by the bi-directional coupling of accurate robot localization and navigation together with strategic environment manipulation. However, many existing robot vision and learning tasks such as visual navigation and robot manipulation address only one of these two coupled aspects. To stimulate the pursuit of a generic and adaptive solution, we reasonably simplify mobile construction as a partially observable Markov decision process (POMDP) in 1/2/3D grid worlds and benchmark the performance of a handcrafted policy with basic localization and planning, and state-of-the-art deep reinforcement learning (RL) methods. Our extensive experiments show that the coupling makes this problem very challenging for those methods, and emphasize the need for novel task-specific solutions.

cs.RO↗

arXiv2021

Marangoni Convection-Driven Laser Fountains and Waves on Free Surfaces of Liquids

It is well accepted that an outward Marangoni convection from a low surface tension region will make the surface depressed. Here, we report that this established perception is only valid for thin liquid films. Using surface laser heating, we show that in deep liquids a laser beam actually pulls up the fluid above the free surface generating fountains with different shapes. Whereas with decreasing liquid depth a transition from fountain to indentation with fountain in-indentation is observed. Further, high-speed imaging reveals a transient surface process before steady elevation is formed, and this dynamic deformation is subsequently utilized to resonantly excite giant surface waves by a modulated laser beam. Computational fluid dynamics models reveal the underlying flow patterns and quantify the depth-dependent and time-resolved surface deformations. Our discoveries and techniques have upended the century-old perception and opened up a new regime of interdisciplinary research and applications of Marangoni-induced interface phenomena and optocapillary fluidic surfaces-the control of fluids with light.

physics.flu-dyn↗

arXiv2021

Generating Diverse Structure for Image Inpainting With Hierarchical VQ-VAE

Given an incomplete image without additional constraint, image inpainting natively allows for multiple solutions as long as they appear plausible. Recently, multiplesolution inpainting methods have been proposed and shown the potential of generating diverse results. However, these methods have difficulty in ensuring the quality of each solution, e.g. they produce distorted structure and/or blurry texture. We propose a two-stage model for diverse inpainting, where the first stage generates multiple coarse results each of which has a different structure, and the second stage refines each coarse result separately by augmenting texture. The proposed model is inspired by the hierarchical vector quantized variational auto-encoder (VQ-VAE), whose hierarchical architecture isentangles structural and textural information. In addition, the vector quantization in VQVAE enables autoregressive modeling of the discrete distribution over the structural information. Sampling from the distribution can easily generate diverse and high-quality structures, making up the first stage of our model. In the second stage, we propose a structural attention module inside the texture generation network, where the module utilizes the structural information to capture distant correlations. We further reuse the VQ-VAE to calculate two feature losses, which help improve structure coherence and texture realism, respectively. Experimental results on CelebA-HQ, Places2, and ImageNet datasets show that our method not only enhances the diversity of the inpainting solutions but also improves the visual quality of the generated multiple images. Code and models are available at: https://github.com/USTC-JialunPeng/Diverse-Structure-Inpainting.

cs.CV↗

arXiv2021

An efficient Quasi-Newton method for nonlinear inverse problems via learned singular values

Solving complex optimization problems in engineering and the physical sciences requires repetitive computation of multi-dimensional function derivatives. Commonly, this requires computationally-demanding numerical differentiation such as perturbation techniques, which ultimately limits the use for time-sensitive applications. In particular, in nonlinear inverse problems Gauss-Newton methods are used that require iterative updates to be computed from the Jacobian. Computationally more efficient alternatives are Quasi-Newton methods, where the repeated computation of the Jacobian is replaced by an approximate update. Here we present a highly efficient data-driven Quasi-Newton method applicable to nonlinear inverse problems. We achieve this, by using the singular value decomposition and learning a mapping from model outputs to the singular values to compute the updated Jacobian. This enables a speed-up expected of Quasi-Newton methods without accumulating roundoff errors, enabling time-critical applications and allowing for flexible incorporation of prior knowledge necessary to solve ill-posed problems. We present results for the highly non-linear inverse problem of electrical impedance tomography with experimental data.

math.NA↗

arXiv2021

Synergy Between Semantic Segmentation and Image Denoising via Alternate Boosting

The capability of image semantic segmentation may be deteriorated due to noisy input image, where image denoising prior to segmentation helps. Both image denoising and semantic segmentation have been developed significantly with the advance of deep learning. Thus, we are interested in the synergy between them by using a holistic deep model. We observe that not only denoising helps combat the drop of segmentation accuracy due to noise, but also pixel-wise semantic information boosts the capability of denoising. We then propose a boosting network to perform denoising and segmentation alternately. The proposed network is composed of multiple segmentation and denoising blocks (SDBs), each of which estimates semantic map then uses the map to regularize denoising. Experimental results show that the denoised image quality is improved substantially and the segmentation accuracy is improved to close to that of clean images. Our code and models will be made publicly available.

cs.CV↗