arXiv subjects

Chen Chen

Publications and source records attributed to Chen Chen.

At least 883 records · Page 49Linked to original sources

arXiv2021

Enhanced 3D Human Pose Estimation from Videos by using Attention-Based Neural Network with Dilated Convolutions

The attention mechanism provides a sequential prediction framework for learning spatial models with enhanced implicit temporal consistency. In this work, we show a systematic design (from 2D to 3D) for how conventional networks and other forms of constraints can be incorporated into the attention framework for learning long-range dependencies for the task of pose estimation. The contribution of this paper is to provide a systematic approach for designing and training of attention-based models for the end-to-end pose estimation, with the flexibility and scalability of arbitrary video sequences as input. We achieve this by adapting temporal receptive field via a multi-scale structure of dilated convolutions. Besides, the proposed architecture can be easily adapted to a causal model enabling real-time performance. Any off-the-shelf 2D pose estimation systems, e.g. Mocap libraries, can be easily integrated in an ad-hoc fashion. Our method achieves the state-of-the-art performance and outperforms existing methods by reducing the mean per joint position error to 33.4 mm on Human3.6M dataset.

cs.CV↗

arXiv2021

Foresee then Evaluate: Decomposing Value Estimation with Latent Future Prediction

Value function is the central notion of Reinforcement Learning (RL). Value estimation, especially with function approximation, can be challenging since it involves the stochasticity of environmental dynamics and reward signals that can be sparse and delayed in some cases. A typical model-free RL algorithm usually estimates the values of a policy by Temporal Difference (TD) or Monte Carlo (MC) algorithms directly from rewards, without explicitly taking dynamics into consideration. In this paper, we propose Value Decomposition with Future Prediction (VDFP), providing an explicit two-step understanding of the value estimation process: 1) first foresee the latent future, 2) and then evaluate it. We analytically decompose the value function into a latent future dynamics part and a policy-independent trajectory return part, inducing a way to model latent dynamics and returns separately in value estimation. Further, we derive a practical deep RL algorithm, consisting of a convolutional model to learn compact trajectory representation from past experiences, a conditional variational auto-encoder to predict the latent future dynamics and a convex return model that evaluates trajectory representation. In experiments, we empirically demonstrate the effectiveness of our approach for both off-policy and on-policy RL in several OpenAI Gym continuous control tasks as well as a few challenging variants with delayed reward.

cs.LG↗

arXiv2021

Addressing Action Oscillations through Learning Policy Inertia

Deep reinforcement learning (DRL) algorithms have been demonstrated to be effective in a wide range of challenging decision making and control tasks. However, these methods typically suffer from severe action oscillations in particular in discrete action setting, which means that agents select different actions within consecutive steps even though states only slightly differ. This issue is often neglected since the policy is usually evaluated by its cumulative rewards only. Action oscillation strongly affects the user experience and can even cause serious potential security menace especially in real-world domains with the main concern of safety, such as autonomous driving. To this end, we introduce Policy Inertia Controller (PIC) which serves as a generic plug-in framework to off-the-shelf DRL algorithms, to enables adaptive trade-off between the optimality and smoothness of the learned policy in a formal way. We propose Nested Policy Iteration as a general training algorithm for PIC-augmented policy which ensures monotonically non-decreasing updates under some mild conditions. Further, we derive a practical DRL algorithm, namely Nested Soft Actor-Critic. Experiments on a collection of autonomous driving tasks and several Atari games suggest that our approach demonstrates substantial oscillation reduction in comparison to a range of commonly adopted baselines with almost no performance degradation.

cs.LG↗

arXiv2021

Direct observation of layer-stacking and oriented wrinkles in multilayer hexagonal boron nitride

Hexagonal boron nitride (h-BN) has long been recognized as an ideal substrate for electronic devices due to its dangling-bond-free surface, insulating nature and thermal/chemical stability. Therefore, to analyse the lattice structure and orientation of h-BN crystals becomes important. Here, the stacking order and wrinkles of h-BN are investigated by transmission electron microscopy (TEM). It is experimentally confirmed that the layers in the h-BN flakes are arranged in the AA' stacking. The wrinkles in a form of threefold network throughout the h-BN crystal are oriented along the armchair direction, and their formation mechanism was further explored by molecular dynamics simulations. Our findings provide a deep insight about the microstructure of h-BN and shed light on the structural design/electronic modulations of two-dimensional crystals.

cond-mat.mtrl-sci↗

arXiv2021

Robust Trajectory-Constrained Frequency Control for Microgrids Considering Model Linearization Error

The capability to switch between grid-connected and islanded modes has promoted adoption of microgrid technology for powering remote locations. Stabilizing frequency during the islanding event, however, is a challenging control task, particularly under high penetration of converter-interfaced sources. In this paper, a numerical optimal control (NOC)-based control synthesis methodology is proposed for preparedness of microgrid islanding that ensure guaranteed performance. The key feature of the proposed paradigm is near real-time centralized scheduling for real-time decentralized executing. For tractable computation, linearized models are used in the problem formulation. To accommodate the linearization errors, interval analysis is employed to compute linearization-induced uncertainty as numerical intervals so that the NOC problem can be formulated into a robust mixed-integer linear program. The proposed control is verified on the full nonlinear model in Simulink. The simulation results shown effectiveness of the proposed control paradigm and the necessity of considering linearization-induced uncertainty.

eess.SY↗

arXiv2021

Moiré Imaging in Twisted Bilayer Graphene Aligned on Hexagonal Boron Nitride

Moiré superlattices (MSL) formed in angle-aligned bilayers of van der Waals materials have become a promising platform to realize novel two-dimensional electronic states. Angle-aligned trilayer structures can form two sets of MSLs which could potentially interfere with each other. In this work, we directly image the moiré patterns in both monolayer graphene aligned on hBN and twisted bilayer graphene aligned on hBN, using combined scanning microwave impedance microscopy and conductive atomic force microscopy. Correlation of the two techniques reveals the contrast mechanism for the achieved ultrahigh spatial resolution (<2 nm). We observe two sets of MSLs with different periodicities in the trilayer stack. The smaller MSL breaks the 6-fold rotational symmetry and exhibits abrupt discontinuities at the boundaries of the larger MSL. Using a rigid atomic-stacking model, we demonstrate that the hBN layer considerably modifies the MSL of twisted bilayer graphene. We further analyze its effect on the reciprocal space spectrum of the dual-moiré system.

cond-mat.mes-hall↗

arXiv2021

A Dataset and Benchmark for Malaria Life-Cycle Classification in Thin Blood Smear Images

Malaria microscopy, microscopic examination of stained blood slides to detect parasite Plasmodium, is considered to be a gold-standard for detecting life-threatening disease malaria. Detecting the plasmodium parasite requires a skilled examiner and may take up to 10 to 15 minutes to completely go through the whole slide. Due to a lack of skilled medical professionals in the underdeveloped or resource deficient regions, many cases go misdiagnosed; resulting in unavoidable complications and/or undue medication. We propose to complement the medical professionals by creating a deep learning-based method to automatically detect (localize) the plasmodium parasites in the photograph of stained film. To handle the unbalanced nature of the dataset, we adopt a two-stage approach. Where the first stage is trained to detect blood cells and classify them into just healthy or infected. The second stage is trained to classify each detected cell further into the life-cycle stage. To facilitate the research in machine learning-based malaria microscopy, we introduce a new large scale microscopic image malaria dataset. Thirty-eight thousand cells are tagged from the 345 microscopic images of different Giemsa-stained slides of blood samples. Extensive experimentation is performed using different CNN backbones including VGG, DenseNet, and ResNet on this dataset. Our experiments and analysis reveal that the two-stage approach works better than the one-stage approach for malaria detection. To ensure the usability of our approach, we have also developed a mobile app that will be used by local hospitals for investigation and educational purposes. The dataset, its annotations, and implementation codes will be released upon publication of the paper.

eess.IV↗

arXiv2021

Guided Interpolation for Adversarial Training

To enhance adversarial robustness, adversarial training learns deep neural networks on the adversarial variants generated by their natural data. However, as the training progresses, the training data becomes less and less attackable, undermining the robustness enhancement. A straightforward remedy is to incorporate more training data, but sometimes incurring an unaffordable cost. In this paper, to mitigate this issue, we propose the guided interpolation framework (GIF): in each epoch, the GIF employs the previous epoch's meta information to guide the data's interpolation. Compared with the vanilla mixup, the GIF can provide a higher ratio of attackable data, which is beneficial to the robustness enhancement; it meanwhile mitigates the model's linear behavior between classes, where the linear behavior is favorable to generalization but not to the robustness. As a result, the GIF encourages the model to predict invariantly in the cluster of each class. Experiments demonstrate that the GIF can indeed enhance adversarial robustness on various adversarial training methods and various datasets.

cs.LG↗

arXiv2021

Form Factors of the Nucleon Axial Current

A symmetry-preserving Poincaré-covariant quark+diquark Faddeev equation treatment of the nucleon is used to deliver parameter-free predictions for the nucleon's axial and induced pseudoscalar form factors, $G_A$ and $G_P$, respectively. The result for $G_A$ can reliably be represented by a dipole form factor characterised by an axial charge $g_A=G_A(0)=1.25(3)$ and a mass-scale $M_A = 1.23(3) m_N$, where $m_N$ is the nucleon mass; and regarding $G_P$, the induced pseudoscalar charge $g_p^\ast = 8.80(23)$, the ratio $g_p^\ast/g_A = 7.04(22)$, and the pion pole dominance Ansatz is found to provide a reliable estimate of the directly computed result. The ratio of flavour-separated quark axial charges is also calculated: $g_A^d/g_A^u=-0.16(2)$. This value expresses a marked suppression of the size of the $d$-quark component relative to that found in nonrelativistic quark models and owes to the presence of strong diquark correlations in the nucleon Faddeev wave function -- both scalar and axial-vector, with the scalar diquark being dominant. The predicted form for $G_A$ should provide a sound foundation for analyses of the neutrino-nucleus and antineutrino-nucleus cross-sections that are relevant to modern accelerator neutrino experiments.

hep-ph↗

arXiv2021

A Catalog of High-Velocity CIV Mini-BALs in the VLT-UVES and Keck-HIRES Archives

We present a catalog of high-velocity CIV $λ$ 1548,1551 mini-Broad Absorption Lines (mini-BALs) in the archives of the VLT-UVES and Keck-HIRES spectrographs. We identify high-velocity CIV mini-BALs based on smooth rounded BAL-like profiles with velocity blueshifts $<$ $-$4000 km/s and widths in the range 70 $\lesssim$ FWHM(1548) $\lesssim$ 2000 km/s (for $λ$1548 alone). We find 105 mini-BALs in 44 quasars from a total sample of 638 quasars. The fraction of quasars with at least one mini-BAL meeting our criteria is roughly $\sim9$% after correcting for incomplete velocity coverage. However, the numbers of systems rise sharply at lower velocities and narrower FWHMs, suggesting that many outflow lines are missed by our study. All of the systems are highly ionized based on the strong presence of NV and OVI and/or the absence of SiII and CII when within the wavelength coverage. Two of the mini-BAL systems in our catalog, plus three others at smaller velocity shifts, have PV $λ$1118,1128 absorption indicating highly saturated CIV absorption and total hydrogen column densities $\gtrsim 10^{22}$ cm$^{-3}$. Most of the mini-BALs are confirmed to have optical depths $\gtrsim$1 with partial covering of the quasar continuum source. The covering fractions are as small as 0.06 in CIV and 0.03 in SiIV , corresponding to outflow absorbing structures $<0.002$ pc across. When multiple lines are measured, the lines of less abundant ions tend to have narrower profiles and smaller covering fractions indicative of inhomogeneous absorbers where higher column densities occur in smaller clumps. This picture might extend to BAL outflows if the broader and generally deeper BALs form in either the largest clumps or collections of many mini-BAL-like clumps that blend together in observed quasar spectra.

astro-ph.GA↗

arXiv2021

REMIX: Efficient Range Query for LSM-trees

LSM-tree based key-value (KV) stores organize data in a multi-level structure for high-speed writes. Range queries on traditional LSM-trees must seek and sort-merge data from multiple table files on the fly, which is expensive and often leads to mediocre read performance. To improve range query efficiency on LSM-trees, we introduce a space-efficient KV index data structure, named REMIX, that records a globally sorted view of KV data spanning multiple table files. A range query on multiple REMIX-indexed data files can quickly locate the target key using a binary search, and retrieve subsequent keys in sorted order without key comparisons. We build RemixDB, an LSM-tree based KV-store that adopts a write-efficient compaction strategy and employs REMIXes for fast point and range queries. Experimental results show that REMIXes can substantially improve range query performance in a write-optimized LSM-tree based KV-store.

cs.DB↗

arXiv2021

A Transferable Anti-Forensic Attack on Forensic CNNs Using A Generative Adversarial Network

With the development of deep learning, convolutional neural networks (CNNs) have become widely used in multimedia forensics for tasks such as detecting and identifying image forgeries. Meanwhile, anti-forensic attacks have been developed to fool these CNN-based forensic algorithms. Previous anti-forensic attacks often were designed to remove forgery traces left by a single manipulation operation as opposed to a set of manipulations. Additionally, recent research has shown that existing anti-forensic attacks against forensic CNNs have poor transferability, i.e. they are unable to fool other forensic CNNs that were not explicitly used during training. In this paper, we propose a new anti-forensic attack framework designed to remove forensic traces left by a variety of manipulation operations. This attack is transferable, i.e. it can be used to attack forensic CNNs are unknown to the attacker, and it introduces only minimal distortions that are imperceptible to human eyes. Our proposed attack utilizes a generative adversarial network (GAN) to build a generator that can attack color images of any size. We achieve attack transferability through the use of a new training strategy and loss function. We conduct extensive experiment to demonstrate that our attack can fool many state-of-art forensic CNNs with varying levels of knowledge available to the attacker.

cs.CV↗

arXiv2020

The Hopping Mechanism of the Hydrated Excess Proton and Its Contribution to Proton Diffusion in Water

In this work a series of analyses are performed on ab initio molecular dynamics (AIMD) simulations of a hydrated excess proton in water to quantify the relative occurrence of concerted hopping events and rattling events, and thus to further elucidate the hopping mechanism of proton transport in water. Contrary to results reported in certain earlier papers, the new analysis finds that concerted hopping events do occur in all simulations, but that the majority of events are the product of proton rattling, where the excess proton will rattle between two or more waters. The results are consistent with the proposed special-pair dance model of the hydrated excess proton, wherein the acceptor water molecule for the proton transfer will quickly change (resonate between three equivalent special pairs) until a decisive proton hop occurs. To remove the misleading effect of simple rattling, a filter was applied to the trajectory such that hopping events that were followed by back hops to the original water are not counted. A steep reduction in the number of multiple hopping events is found when the filter is applied, suggesting that many multiple hopping events that occur in the unfiltered trajectory are largely the product of rattling, contrary to prior suggestions. Comparing the continuous correlation function of the filtered and unfiltered trajectories, we find agreement with experimental values for the proton hopping time and Eigen-Zundel interconversion time, respectively.

physics.chem-ph↗

arXiv2020

Towards Effective Context for Meta-Reinforcement Learning: an Approach based on Contrastive Learning

Context, the embedding of previous collected trajectories, is a powerful construct for Meta-Reinforcement Learning (Meta-RL) algorithms. By conditioning on an effective context, Meta-RL policies can easily generalize to new tasks within a few adaptation steps. We argue that improving the quality of context involves answering two questions: 1. How to train a compact and sufficient encoder that can embed the task-specific information contained in prior trajectories? 2. How to collect informative trajectories of which the corresponding context reflects the specification of tasks? To this end, we propose a novel Meta-RL framework called CCM (Contrastive learning augmented Context-based Meta-RL). We first focus on the contrastive nature behind different tasks and leverage it to train a compact and sufficient context encoder. Further, we train a separate exploration policy and theoretically derive a new information-gain-based objective which aims to collect informative trajectories in a few steps. Empirically, we evaluate our approaches on common benchmarks as well as several complex sparse-reward environments. The experimental results show that CCM outperforms state-of-the-art algorithms by addressing previously mentioned problems respectively.

cs.LG↗

arXiv2020

Semi-Dynamic Load Balancing: Efficient Distributed Learning in Non-Dedicated Environments

Machine learning (ML) models are increasingly trained in clusters with non-dedicated workers possessing heterogeneous resources. In such scenarios, model training efficiency can be negatively affected by stragglers -- workers that run much slower than others. Efficient model training requires eliminating such stragglers, yet for modern ML workloads, existing load balancing strategies are inefficient and even infeasible. In this paper, we propose a novel strategy called semi-dynamic load balancing to eliminate stragglers of distributed ML workloads. The key insight is that ML workers shall be load-balanced at iteration boundaries, being non-intrusive to intra-iteration execution. We develop LB-BSP based on such an insight, which is an integrated worker coordination mechanism that adapts workers' load to their instantaneous processing capabilities by right-sizing the sample batches at the synchronization barriers. We have custom-designed the batch sizing algorithm respectively for CPU and GPU clusters based on their own characteristics. LB-BSP has been implemented as a Python module for ML frameworks like TensorFlow and PyTorch. Our EC2 deployment confirms that LB-BSP is practical, effective and light-weight, and is able to accelerating distributed training by up to $54\%$.

cs.DC↗

arXiv2020

Revisiting Street-to-Aerial View Image Geo-localization and Orientation Estimation

Street-to-aerial image geo-localization, which matches a query street-view image to the GPS-tagged aerial images in a reference set, has attracted increasing attention recently. In this paper, we revisit this problem and point out the ignored issue about image alignment information. We show that the performance of a simple Siamese network is highly dependent on the alignment setting and the comparison of previous works can be unfair if they have different assumptions. Instead of focusing on the feature extraction under the alignment assumption, we show that improvements in metric learning techniques significantly boost the performance regardless of the alignment. Without leveraging the alignment information, our pipeline outperforms previous works on both panorama and cropped datasets. Furthermore, we conduct visualization to help understand the learned model and the effect of alignment information using Grad-CAM. With our discovery on the approximate rotation-invariant activation maps, we propose a novel method to estimate the orientation/alignment between a pair of cross-view images with unknown alignment information. It achieves state-of-the-art results on the CVUSA dataset.

cs.CV↗

arXiv2020

Approximating Trajectory Constraints with Machine Learning -- Microgrid Islanding with Frequency Constraints

In this paper, we introduce a deep learning aided constraint encoding method to tackle the frequency-constraint microgrid scheduling problem. The nonlinear function between system operating condition and frequency nadir is approximated by using a neural network, which admits an exact mixed-integer formulation (MIP). This formulation is then integrated with the scheduling problem to encode the frequency constraint. With the stronger representation power of the neural network, the resulting commands can ensure adequate frequency response in a realistic setting in addition to islanding success. The proposed method is validated on a modified 33-node system. Successful islanding with a secure response is simulated under the scheduled commands using a detailed three-phase model in Simulink. The advantages of our model are particularly remarkable when the inertia emulation functions from wind turbine generators are considered.

eess.SY↗

arXiv2020

VEGA: Towards an End-to-End Configurable AutoML Pipeline

Automated Machine Learning (AutoML) is an important industrial solution for automatic discovery and deployment of the machine learning models. However, designing an integrated AutoML system faces four great challenges of configurability, scalability, integrability, and platform diversity. In this work, we present VEGA, an efficient and comprehensive AutoML framework that is compatible and optimized for multiple hardware platforms. a) The VEGA pipeline integrates various modules of AutoML, including Neural Architecture Search (NAS), Hyperparameter Optimization (HPO), Auto Data Augmentation, Model Compression, and Fully Train. b) To support a variety of search algorithms and tasks, we design a novel fine-grained search space and its description language to enable easy adaptation to different search algorithms and tasks. c) We abstract the common components of deep learning frameworks into a unified interface. VEGA can be executed with multiple back-ends and hardwares. Extensive benchmark experiments on multiple tasks demonstrate that VEGA can improve the existing AutoML algorithms and discover new high-performance models against SOTA methods, e.g. the searched DNet model zoo for Ascend 10x faster than EfficientNet-B5 and 9.2x faster than RegNetX-32GF on ImageNet. VEGA is open-sourced at https://github.com/huawei-noah/vega.

cs.CV↗