arXiv subjects

Dong Wang

Publications and source records attributed to Dong Wang.

At least 559 records · Page 31Linked to original sources

arXiv2021

A Dataset And Benchmark Of Underwater Object Detection For Robot Picking

Underwater object detection for robot picking has attracted a lot of interest. However, it is still an unsolved problem due to several challenges. We take steps towards making it more realistic by addressing the following challenges. Firstly, the currently available datasets basically lack the test set annotations, causing researchers must compare their method with other SOTAs on a self-divided test set (from the training set). Training other methods lead to an increase in workload and different researchers divide different datasets, resulting there is no unified benchmark to compare the performance of different algorithms. Secondly, these datasets also have other shortcomings, e.g., too many similar images or incomplete labels. Towards these challenges we introduce a dataset, Detecting Underwater Objects (DUO), and a corresponding benchmark, based on the collection and re-annotation of all relevant datasets. DUO contains a collection of diverse underwater images with more rational annotations. The corresponding benchmark provides indicators of both efficiency and accuracy of SOTAs (under the MMDtection framework) for academic research and industrial applications, where JETSON AGX XAVIER is used to assess detector speed to simulate the robot-embedded environment.

cs.CV↗

arXiv2021

Automatic Patch Linkage Detection in Code Review Using TextualContent and File Location Features

Context: Contemporary code review tools are a popular choice for software quality assurance. Using these tools, reviewers are able to post a linkage between two patches during a review discussion. Large development teams that use a review-then-commit model risk being unaware of these linkages. Objective: Our objective is to first explore how patch linkage impacts the review process. We then propose and evaluate models that detect patch linkage based on realistic time intervals. Method: First, we carry out an exploratory study on three open source projects to conduct linkage impact analysis using 942 manually classified linkages. Second, we propose two techniques using textual and file location similarity to build detection models and evaluate their performance. Results: The study provides evidence of latency in the linkage notification. We show that a patch with the Alternative Solution linkage (i.e., patches that implement similar functionality)undergoes a quicker review and avoids additional revisions after the team has been notified, compared to other linkage types. Our detection model experiments show promising recall rates for the Alternative Solution linkage (from 32% to 95%), but precision has room for improvement. Conclusion: Patch linkage detection is promising, with likely improvements if the practice of posting linkages becomes more prevalent. From our implications, this paper lays the groundwork for future research on how to increase patch linkage awareness to facilitate efficient reviews.

cs.SE↗

arXiv2021

Topological error correction with a Gaussian cluster state

Topological error correction provides an effective method to correct errors in quantum computation. It allows quantum computation to be implemented with higher error threshold and high tolerating loss rates. We present a topological a error correction scheme with continuous variables based on an eight-partite Gaussian cluster state. We show that topological quantum correlation between two modes can be protected against a single quadrature phase displacement error occurring on any mode and some of two errors occurring on two modes. More interestingly, some cases of errors occurring on three modes can also be recognised and corrected, which is different from the topological error correction with discrete variables. We show that the final error rate after correction can be further reduced if the modes are subjected to identical errors occurring on all modes with equal probability. The presented results provide a feasible scheme for topological error correction with continuous variables and it can be experimentally demonstrated with a Gaussian cluster state.

quant-ph↗

arXiv2021

Bridging particle deformability and collective response in soft solids

Soft, amorphous solids such as tissues, foams, and emulsions are composed of deformable particles. However, the effect of single-particle deformability on the collective behavior of soft solids is still poorly understood. We perform numerical simulations of two-dimensional jammed packings of explicitly deformable particles to study the mechanical response of model soft solids. We find that jammed packings of deformable particles with excess shape degrees of freedom possess low-frequency quartic vibrational modes that stabilize the packings even though they possess fewer interparticle contacts than the nominal isostatic value. Adding intra-particle constraints can rigidify the particles, but these particles undergo a buckling transition and gain an effective shape degree of freedom when their preferred perimeter is above a threshold value. We find that the mechanical response of jammed packings of deformable particles with shape degrees of freedom differs significantly from that of jammed packings of rigid-shape particles, which emphasizes the importance of particle deformability in modelling soft solids.

cond-mat.soft↗

arXiv2021

A Characteristic Function-based Algorithm for Geodesic Active Contours

Active contour models have been widely used in image segmentation, and the level set method (LSM) is the most popular approach for solving the models, via implicitly representing the contour by a level set function. However, the LSM suffers from high computational burden and numerical instability, requiring additional regularization terms or re-initialization techniques. In this paper, we use characteristic functions to implicitly represent the contours, propose a new representation to the geodesic active contours and derive an efficient algorithm termed as the iterative convolution-thresholding method (ICTM). Compared to the LSM, the ICTM is simpler and much more efficient. In addition, the ICTM enjoys most desired features of the level set-based methods. Extensive experiments, on 2D synthetic, 2D ultrasound, 3D CT, and 3D MR images for nodule, organ and lesion segmentation, demonstrate that the proposed method not only obtains comparable or even better segmentation results (compared to the LSM) but also achieves significant acceleration.

cs.CV↗

arXiv2021

Privacy-preserving Travel Time Prediction with Uncertainty Using GPS Trace Data

The rapid growth of GPS technology and mobile devices has led to a massive accumulation of location data, bringing considerable benefits to individuals and society. One of the major usages of such data is travel time prediction, a typical service provided by GPS navigation devices and apps. Meanwhile, the constant collection and analysis of the individual location data also pose unprecedented privacy threats. We leverage the notion of geo-indistinguishability, an extension of differential privacy to the location privacy setting, and propose a procedure for privacy-preserving travel time prediction without collecting actual individual GPS trace data. We propose new concepts to examine the impact of geo-indistinguishability-based sanitization on the usefulness of GPS traces and provide analytical and experimental utility analysis for privacy-preserving travel time prediction. We also propose new metrics to measure the adversary error in learning individual GPS traces from the collected sanitized data. Our experiment results suggest that the proposed procedure provides travel time prediction with satisfactory accuracy at reasonably small privacy costs.

cs.CR↗

arXiv2021

Proactive Pseudo-Intervention: Causally Informed Contrastive Learning For Interpretable Vision Models

Deep neural networks excel at comprehending complex visual signals, delivering on par or even superior performance to that of human experts. However, ad-hoc visual explanations of model decisions often reveal an alarming level of reliance on exploiting non-causal visual cues that strongly correlate with the target label in training data. As such, deep neural nets suffer compromised generalization to novel inputs collected from different sources, and the reverse engineering of their decision rules offers limited interpretability. To overcome these limitations, we present a novel contrastive learning strategy called {\it Proactive Pseudo-Intervention} (PPI) that leverages proactive interventions to guard against image features with no causal relevance. We also devise a novel causally informed salience mapping module to identify key image pixels to intervene, and show it greatly facilitates model interpretability. To demonstrate the utility of our proposals, we benchmark on both standard natural images and challenging medical image datasets. PPI-enhanced models consistently deliver superior performance relative to competing solutions, especially on out-of-domain predictions and data integration from heterogeneous sources. Further, our causally trained saliency maps are more succinct and meaningful relative to their non-causal counterparts.

cs.CV↗

arXiv2021

LightTrack: Finding Lightweight Neural Networks for Object Tracking via One-Shot Architecture Search

Object tracking has achieved significant progress over the past few years. However, state-of-the-art trackers become increasingly heavy and expensive, which limits their deployments in resource-constrained applications. In this work, we present LightTrack, which uses neural architecture search (NAS) to design more lightweight and efficient object trackers. Comprehensive experiments show that our LightTrack is effective. It can find trackers that achieve superior performance compared to handcrafted SOTA trackers, such as SiamRPN++ and Ocean, while using much fewer model Flops and parameters. Moreover, when deployed on resource-constrained mobile chipsets, the discovered trackers run much faster. For example, on Snapdragon 845 Adreno GPU, LightTrack runs $12\times$ faster than Ocean, while using $13\times$ fewer parameters and $38\times$ fewer Flops. Such improvements might narrow the gap between academic models and industrial deployments in object tracking task. LightTrack is released at https://github.com/researchmm/LightTrack.

cs.CV↗

arXiv2021

Eigenvector distribution in the critical regime of BBP transition

In this paper, we study the random matrix model of Gaussian Unitary Ensemble (GUE) with fixed-rank (aka spiked) external source. We will focus on the critical regime of the Baik-Ben Arous-Péché (BBP) phase transition and establish the distribution of the eigenvectors associated with the leading eigenvalues. The distribution is given in terms of a determinantal point process with extended Airy kernel. Our result can be regarded as an eigenvector counterpart of the BBP eigenvalue phase transition (arXiv:math/0403022). The derivation of the distribution makes use of the recently re-discovered eigenvector-eigenvalue identity, together with the determinantal point process representation of the GUE minor process with external source.

math.PR↗

arXiv2021

Nonmonotonic confining potential and eigenvalue density transition for generalized random matrix model

We consider several limiting cases of the joint probability distribution for a random matrix ensemble with an additional interaction term controlled by an exponent $γ$ (called the $γ$-ensembles). The effective potential, which is essentially the single-particle confining potential for an equivalent ensemble with $γ=1$ (called the Muttalib-Borodin ensemble), is a crucial quantity defined in solution to the Riemann-Hilbert problem associated with the $γ$-ensembles. It enables us to numerically compute the eigenvalue density of $γ$-ensembles for all $γ> 0$. We show that one important effect of the two-particle interaction parameter $γ$ is to generate or enhance the non-monotonicity in the effective single-particle potential. For suitable choices of the initial single-particle potentials, reducing $γ$ can lead to a large non-monotonicity in the effective potential, which in turn leads to significant changes in the density of eigenvalues. For a disordered conductor, this corresponds to a systematic decrease in the conductance with increasing disorder. This suggests that appropriate models of $γ$-ensembles can be used as a possible framework to study the effects of disorder on the distribution of conductances.

cond-mat.dis-nn↗

arXiv2021

Can We Benchmark Code Review Studies? A Systematic Mapping Study of Methodology, Dataset, and Metric

Code Review (CR) is the cornerstone for software quality assurance and a crucial practice for software development. As CR research matures, it can be difficult to keep track of the best practices and state-of-the-art in methodology, dataset, and metric. This paper investigates the potential of benchmarking by collecting methodology, dataset, and metric of CR studies. A systematic mapping study was conducted. A total of 112 studies from 19,847 papers published in high-impact venues between the years 2011 and 2019 were selected and analyzed. First, we find that empirical evaluation is the most common methodology (65% of papers), with solution and experience being the least common methodology. Second, we highlight 50% of papers that use the quantitative method or mixed-method have the potential for replicability. Third, we identify 457 metrics that are grouped into sixteen core metric sets, applied to nine Software Engineering topics, showing different research topics tend to use specific metric sets. We conclude that at this stage, we cannot benchmark CR studies. Nevertheless, a common benchmark will facilitate new researchers, including experts from other fields, to innovate new techniques and build on top of already established methodologies. A full replication is available at https://naist-se.github.io/code-review/.

cs.SE↗

arXiv2021

Alpha-Refine: Boosting Tracking Performance by Precise Bounding Box Estimation

In recent years, the multiple-stage strategy has become a popular trend for visual tracking. This strategy first utilizes a base tracker to coarsely locate the target and then exploits a refinement module to obtain more accurate results. However, existing refinement modules suffer from the limited transferability and precision. In this work, we propose a novel, flexible and accurate refinement module called Alpha-Refine, which exploits a precise pixel-wise correlation layer together with a spatial-aware non-local layer to fuse features and can predict three complementary outputs: bounding box, corners and mask. To wisely choose the most adequate output, we also design a light-weight branch selector module. We apply the proposed Alpha-Refine module to five famous and state-of-the-art base trackers: DiMP, ATOM, SiamRPN++, RTMDNet and ECO. The comprehensive experiments on TrackingNet, LaSOT and VOT2018 benchmarks demonstrate that our approach significantly improves the tracking performance in comparison with other existing refinement methods. The source codes will be available at https://github.com/MasterBin-IIAU/AlphaRefine.

cs.CV↗

arXiv2021

Learning Spatio-Temporal Transformer for Visual Tracking

In this paper, we present a new tracking architecture with an encoder-decoder transformer as the key component. The encoder models the global spatio-temporal feature dependencies between target objects and search regions, while the decoder learns a query embedding to predict the spatial positions of the target objects. Our method casts object tracking as a direct bounding box prediction problem, without using any proposals or predefined anchors. With the encoder-decoder transformer, the prediction of objects just uses a simple fully-convolutional network, which estimates the corners of objects directly. The whole method is end-to-end, does not need any postprocessing steps such as cosine window and bounding box smoothing, thus largely simplifying existing tracking pipelines. The proposed tracker achieves state-of-the-art performance on five challenging short-term and long-term benchmarks, while running at real-time speed, being 6x faster than Siam R-CNN. Code and models are open-sourced at https://github.com/researchmm/Stark.

cs.CV↗

arXiv2021

Wasserstein Contrastive Representation Distillation

The primary goal of knowledge distillation (KD) is to encapsulate the information of a model learned from a teacher network into a student network, with the latter being more compact than the former. Existing work, e.g., using Kullback-Leibler divergence for distillation, may fail to capture important structural knowledge in the teacher network and often lacks the ability for feature generalization, particularly in situations when teacher and student are built to address different classification tasks. We propose Wasserstein Contrastive Representation Distillation (WCoRD), which leverages both primal and dual forms of Wasserstein distance for KD. The dual form is used for global knowledge transfer, yielding a contrastive learning objective that maximizes the lower bound of mutual information between the teacher and the student networks. The primal form is used for local contrastive knowledge transfer within a mini-batch, effectively matching the distributions of features between the teacher and the student networks. Experiments demonstrate that the proposed WCoRD method outperforms state-of-the-art approaches on privileged information distillation, model compression and cross-modal transfer.

cs.LG↗

arXiv2021

Transformer Tracking

Correlation acts as a critical role in the tracking field, especially in recent popular Siamese-based trackers. The correlation operation is a simple fusion manner to consider the similarity between the template and the search region. However, the correlation operation itself is a local linear matching process, leading to lose semantic information and fall into local optimum easily, which may be the bottleneck of designing high-accuracy tracking algorithms. Is there any better feature fusion method than correlation? To address this issue, inspired by Transformer, this work presents a novel attention-based feature fusion network, which effectively combines the template and search region features solely using attention. Specifically, the proposed method includes an ego-context augment module based on self-attention and a cross-feature augment module based on cross-attention. Finally, we present a Transformer tracking (named TransT) method based on the Siamese-like feature extraction backbone, the designed attention-based fusion mechanism, and the classification and regression head. Experiments show that our TransT achieves very promising results on six challenging datasets, especially on large-scale LaSOT, TrackingNet, and GOT-10k benchmarks. Our tracker runs at approximatively 50 fps on GPU. Code and models are available at https://github.com/chenxin-dlut/TransT.

cs.CV↗

arXiv2021

Adversarial and Contrastive Variational Autoencoder for Sequential Recommendation

Sequential recommendation as an emerging topic has attracted increasing attention due to its important practical significance. Models based on deep learning and attention mechanism have achieved good performance in sequential recommendation. Recently, the generative models based on Variational Autoencoder (VAE) have shown the unique advantage in collaborative filtering. In particular, the sequential VAE model as a recurrent version of VAE can effectively capture temporal dependencies among items in user sequence and perform sequential recommendation. However, VAE-based models suffer from a common limitation that the representational ability of the obtained approximate posterior distribution is limited, resulting in lower quality of generated samples. This is especially true for generating sequences. To solve the above problem, in this work, we propose a novel method called Adversarial and Contrastive Variational Autoencoder (ACVAE) for sequential recommendation. Specifically, we first introduce the adversarial training for sequence generation under the Adversarial Variational Bayes (AVB) framework, which enables our model to generate high-quality latent variables. Then, we employ the contrastive loss. The latent variables will be able to learn more personalized and salient characteristics by minimizing the contrastive loss. Besides, when encoding the sequence, we apply a recurrent and convolutional structure to capture global and local relationships in the sequence. Finally, we conduct extensive experiments on four real-world datasets. The experimental results show that our proposed ACVAE model outperforms other state-of-the-art methods.

cs.IR↗

arXiv2021

FAST-Dynamic-Vision: Detection and Tracking Dynamic Objects with Event and Depth Sensing

The development of aerial autonomy has enabled aerial robots to fly agilely in complex environments. However, dodging fast-moving objects in flight remains a challenge, limiting the further application of unmanned aerial vehicles (UAVs). The bottleneck of solving this problem is the accurate perception of rapid dynamic objects. Recently, event cameras have shown great potential in solving this problem. This paper presents a complete perception system including ego-motion compensation, object detection, and trajectory prediction for fast-moving dynamic objects with low latency and high precision. Firstly, we propose an accurate ego-motion compensation algorithm by considering both rotational and translational motion for more robust object detection. Then, for dynamic object detection, an event camera-based efficient regression algorithm is designed. Finally, we propose an optimizationbased approach that asynchronously fuses event and depth cameras for trajectory prediction. Extensive real-world experiments and benchmarks are performed to validate our framework. Moreover, our code will be released to benefit related researches.

cs.RO↗

arXiv2021

Necessary and sufficient criterion of steering for two-qubit T states

Einstein-Podolsky-Rosen (EPR) steering is the ability that an observer persuades a distant observer to share entanglement by making local measurements. Determining a quantum state is steerable or unsteerable remains an open problem. Here, we derive a new steering inequality with infinite measurements corresponding to an arbitrary two-qubit T state, from consideration of EPR steering inequalities with N projective measurement settings for each side. In fact, the steering inequality is also a sufficient criterion for guaranteering that the T state is unsteerable. Hence, the steering inequality can be viewed as a necessary and sufficient criterion to distinguish whether the T state is steerable or unsteerable. In order to reveal the fact that the set composed of steerable states is the strict subset of the set made up of entangled states, we prove theoretically that all separable T states can not violate the steering inequality. Moreover, we put forward a method to estimate the maximum violation from concurrence for arbitrary two-qubit T states, which indicates that the T state is steerable if its concurrence exceeds 1/4.

quant-ph↗