Search arXiv⌕ Search

arXiv subjects

Yue Wu

Publications and source records attributed to Yue Wu.

At least 361 records · Page 20Linked to original sources
arXiv2019

MMDetection: Open MMLab Detection Toolbox and Benchmark

We present MMDetection, an object detection toolbox that contains a rich set of object detection and instance segmentation methods as well as related components and modules. The toolbox started from a codebase of MMDet team who won the detection track of COCO Challenge 2018. It gradually evolves into a unified platform that covers many popular detection methods and contemporary modules. It not only includes training and inference codes, but also provides weights for more than 200 network models. We believe this toolbox is by far the most complete detection toolbox. In this paper, we introduce the various features of this toolbox. In addition, we also conduct a benchmarking study on different methods, components, and their hyper-parameters. We wish that the toolbox and benchmark could serve the growing research community by providing a flexible toolkit to reimplement existing methods and develop their own new detectors. Code and models are available at https://github.com/open-mmlab/mmdetection. The project is under active development and we will keep this document updated.

cs.CV↗
arXiv2019

Distributed Learning of Decentralized Control Policies for Articulated Mobile Robots

State-of-the-art distributed algorithms for reinforcement learning rely on multiple independent agents, which simultaneously learn in parallel environments while asynchronously updating a common, shared policy. Moreover, decentralized control architectures (e.g., CPGs) can coordinate spatially distributed portions of an articulated robot to achieve system-level objectives. In this work, we investigate the relationship between distributed learning and decentralized control by learning decentralized control policies for the locomotion of articulated robots in challenging environments. To this end, we present an approach that leverages the structure of the asynchronous advantage actor-critic (A3C) algorithm to provide a natural means of learning decentralized control policies on a single articulated robot. Our primary contribution shows individual agents in the A3C algorithm can be defined by independently controlled portions of the robot's body, thus enabling distributed learning on a single robot for efficient hardware implementation. We present results of closed-loop locomotion in unstructured terrains on a snake and a hexapod robot, using decentralized controllers learned offline and online respectively. Preprint of the paper submitted to the IEEE Transactions in Robotics (T-RO) journal in October 2018, and accepted for publication as a regular paper in May 2019.

cs.RO↗
arXiv2019

Large Scale Incremental Learning

Modern machine learning suffers from catastrophic forgetting when learning new classes incrementally. The performance dramatically degrades due to the missing data of old classes. Incremental learning methods have been proposed to retain the knowledge acquired from the old classes, by using knowledge distilling and keeping a few exemplars from the old classes. However, these methods struggle to scale up to a large number of classes. We believe this is because of the combination of two factors: (a) the data imbalance between the old and new classes, and (b) the increasing number of visually similar classes. Distinguishing between an increasing number of visually similar classes is particularly challenging, when the training data is unbalanced. We propose a simple and effective method to address this data imbalance issue. We found that the last fully connected layer has a strong bias towards the new classes, and this bias can be corrected by a linear model. With two bias parameters, our method performs remarkably well on two large datasets: ImageNet (1000 classes) and MS-Celeb-1M (10000 classes), outperforming the state-of-the-art algorithms by 11.1% and 13.2% respectively.

cs.CV↗
arXiv2019

A Note on Random Sampling for Matrix Multiplication

This paper extends the framework of randomised matrix multiplication to a coarser partition and proposes an algorithm as a complement to the classical algorithm, especially when the optimal probability distribution of the latter one is closed to uniform. The new algorithm increases the likelihood of getting a small approximation error in 2-norm and has the squared approximation error in Frobenious norm bounded by that from the classical algorithm.

math.NA↗
arXiv2019

AIRD: Adversarial Learning Framework for Image Repurposing Detection

Image repurposing is a commonly used method for spreading misinformation on social media and online forums, which involves publishing untampered images with modified metadata to create rumors and further propaganda. While manual verification is possible, given vast amounts of verified knowledge available on the internet, the increasing prevalence and ease of this form of semantic manipulation call for the development of robust automatic ways of assessing the semantic integrity of multimedia data. In this paper, we present a novel method for image repurposing detection that is based on the real-world adversarial interplay between a bad actor who repurposes images with counterfeit metadata and a watchdog who verifies the semantic consistency between images and their accompanying metadata, where both players have access to a reference dataset of verified content, which they can use to achieve their goals. The proposed method exhibits state-of-the-art performance on location-identity, subject-identity and painting-artist verification, showing its efficacy across a diverse set of scenarios.

cs.CV↗
arXiv2019

QATM: Quality-Aware Template Matching For Deep Learning

Finding a template in a search image is one of the core problems many computer vision, such as semantic image semantic, image-to-GPS verification \etc. We propose a novel quality-aware template matching method, QATM, which is not only used as a standalone template matching algorithm, but also a trainable layer that can be easily embedded into any deep neural network. Specifically, we assess the quality of a matching pair using soft-ranking among all matching pairs, and thus different matching scenarios such as 1-to-1, 1-to-many, and many-to-many will be all reflected to different values. Our extensive evaluation on classic template matching benchmarks and deep learning tasks demonstrate the effectiveness of QATM. It not only outperforms state-of-the-art template matching methods when used alone, but also largely improves existing deep network solutions.

cs.CV↗
arXiv2019

Moduli Spaces of Unordered $n\ge5$ Points on the Riemann Sphere and Their Singularities

For $n\ge5$, it is well known that the moduli space $\mathfrak{M_{0,\:n}}$ of unordered $n$ points on the Riemann sphere is a quotient space of the Zariski open set $K_n$ of $\mathbb C^{n-3}$ by an $S_n$ action. The stabilizers of this $S_n$ action at certain points of this Zariski open set $K_n$ correspond to the groups fixing the sets of $n$ points on the Riemann sphere. Let $α$ be a subset of $n$ distinct points on the Riemann sphere. We call the group of all linear fractional transformations leaving $α$ invariant the stabilizer of $α$, which is finite by observation. For each non-trivial finite subgroup $G$ of the group ${\rm PSL}(2,{\Bbb C})$ of linear fractional transformations, we give the necessary and sufficient condition for finite subsets of the Riemann sphere under which the stabilizers of them are conjugate to $G$. We also prove that there does exist some finite subset of the Riemann sphere whose stabilizer coincides with $G$. Next we obtain the irreducible decompositions of the representations of the stabilizers on the tangent spaces at the singularities of $\mathfrak{M_{0,\:n}}$. At last, on $\mathfrak{M_{0,\:5}}$ and $\mathfrak{M_{0,\:6}}$, we work out explicitly the singularities and the representations of their stabilizers on the tangent spaces at them.

math.AG↗
arXiv2019

Earthquake Detection in 1-D Time Series Data with Feature Selection and Dictionary Learning

Earthquakes can be detected by matching spatial patterns or phase properties from 1-D seismic waves. Current earthquake detection methods, such as waveform correlation and template matching, have difficulty detecting anomalous earthquakes that are not similar to other earthquakes. In recent years, machine-learning techniques for earthquake detection have been emerging as a new active research direction. In this paper, we develop a novel earthquake detection method based on dictionary learning. Our detection method first generates rich features via signal processing and statistical methods and further employs feature selection techniques to choose features that carry the most significant information. Based on these selected features, we build a dictionary for classifying earthquake events from non-earthquake events. To evaluate the performance of our dictionary-based detection methods, we test our method on a labquake dataset from Penn State University, which contains 3,357,566 time series data points with a 400 MHz sampling rate. 1,000 earthquake events are manually labeled in total, and the length of these earthquake events varies from 74 to 7151 data points. Through comparison to other detection methods, we show that our feature selection and dictionary learning incorporated earthquake detection method achieves an 80.1% prediction accuracy and outperforms the baseline methods in earthquake detection, including Template Matching (TM) and Support Vector Machine (SVM).

physics.geo-ph↗
arXiv2019

A Data-Driven CO2 Leakage Detection Using Seismic Data and Spatial-Temporal Densely Connected Convolutional Neural Networks

In carbon capture and sequestration, developing effective monitoring methods is needed to detect and respond to CO2 leakage. CO2 leakage detection methods rely on geophysical observations and monitoring sensor network. However, traditional methods usually require the development of site-specific physical models and expert interpretation, and the effectiveness of these methods can be limited to different application locations, operational scenarios, and conditions. In this paper, we developed a novel data-driven leakage detection method based on densely connected convolutional neural networks. Our method differs from conventional leakage monitoring methods by directly learning a mapping relationship between seismic data and the CO2 leakage mass. To account for the spatial and temporal characteristics of seismic data, our novel networks architecture combines 1D and 2D convolutional neural networks. To overcome the computational expense of solving optimization problems, we apply a densely-connecting strategy in our network architecture that reduces the number of network parameters. Based on the features generated by our convolutional neural networks, we further incorporate a long short-term memory network to utilize time-sequential information, which further improves the detection accuracy. Finally, we employ our detection method to synthetic seismic datasets generated based on flow simulations of a hypothetical CO2 storage scenario with injection into a partially compartmentalized sandstone storage reservoir. To evaluate method performance, we conducted multiple experiments including a random leakage test, a sequential test, and a robustness test. Numerical results show that our CO2 leakage detection method successfully detects the leakage and accurately predicts the leakage mass, suggesting that it has the potential for application in monitoring of real CO2 storage sites.

physics.geo-ph↗
arXiv2019

InversionNet: A Real-Time and Accurate Full Waveform Inversion with CNNs and continuous CRFs

Full-waveform inversion problems are usually formulated as optimization problems, where the forward-wave propagation operator $f$ maps the subsurface velocity structures to seismic signals. The existing computational methods for solving full-waveform inversion are not only computationally expensive, but also yields low-resolution results because of the ill-posedness and cycle skipping issues of full-waveform inversion. To resolve those issues, we employ machine-learning techniques to solve the full-waveform inversion. Specifically, we focus on applying the convolutional neural network~(CNN) to directly derive the inversion operator $f^{-1}$ so that the velocity structure can be obtained without knowing the forward operator $f$. We build a convolutional neural network with an encoder-decoder structure to model the correspondence from seismic data to subsurface velocity structures. Furthermore, we employ the conditional random field~(CRF) on top of the CNN to generate structural predictions by modeling the interactions between different locations on the velocity model. Our numerical examples using synthetic seismic reflection data show that the propose CNN-CRF model significantly improve the accuracy of the velocity inversion while the computational time is reduced.

eess.SP↗
arXiv2018

Rayleigh fading suppression in one-dimension optical scatters

Highly coherent wave is favorable for applications in which phase retrieval is necessary, yet a high coherent wave is prone to encounter Rayleigh fading phenomenon as it passes through a medium of random scatters. As an exemplary case, phase-sensitive optical time-domain reflectometry (Φ-OTDR) utilizes coherent interference of backscattering light along a fiber to achieve ultra-sensitive acoustic sensing, but sensing locations with fading won't be functional. Apart from the sensing domain, fading is also ubiquitous in optical imaging and wireless telecommunication, therefore it is of great interest. In this paper, we theoretically describe and experimentally verify how the fading phenomena in one-dimension optical scatters will be suppressed with arbitrary number of independent probing channels. We initially theoretically explained why fading would cause severe noise in the demodulated phase of Φ-OTDR; then M-degree summation of incoherent scattered light-waves is studied for the purpose of eliminating fading. Finally, the gain of the retrieved phase signal-to-noise-ratio and its fluctuations were analytically derived and experimentally verified. This work provides a guideline for fading elimination in one-dimension optical scatters, and it also provides insight for optical imaging and wireless telecommunication.

eess.SP↗
arXiv2018

Towards Understanding Learning Representations: To What Extent Do Different Neural Networks Learn the Same Representation

It is widely believed that learning good representations is one of the main reasons for the success of deep neural networks. Although highly intuitive, there is a lack of theory and systematic approach quantitatively characterizing what representations do deep neural networks learn. In this work, we move a tiny step towards a theory and better understanding of the representations. Specifically, we study a simpler problem: How similar are the representations learned by two networks with identical architecture but trained from different initializations. We develop a rigorous theory based on the neuron activation subspace match model. The theory gives a complete characterization of the structure of neuron activation subspace matches, where the core concepts are maximum match and simple match which describe the overall and the finest similarity between sets of neurons in two networks respectively. We also propose efficient algorithms to find the maximum match and simple matches. Finally, we conduct extensive experiments using our algorithms. Experimental results suggest that, surprisingly, representations learned by the same convolutional layers of networks trained from different initializations are not as similar as prevalently expected, at least in terms of subspace match.

cs.LG↗
arXiv2018

Research on CRO's Dilemma In Sapiens Chain: A Game Theory Method

In recent years, blockchain-based techniques have been widely used in cybersecurity, owing to the decentralization, anonymity, credibility and not be tampered properties of the blockchain. As one of the decentralized framework, Sapiens Chain was proposed to protect cybersecurity by scheduling the computational resources dynamically, which were owned by Computational Resources Owners (CROs). However, when CROs in the same pool attack each other, all CROs will earn less. In this paper, we tackle the problem of prisoner's dilemma from the perspective of CROs. We first define a game that a CRO infiltrates another pool and perform an attack. In such game, the honest CRO can control the payoffs and increase its revenue. By simulating this game, we propose to apply Zero Determinant (ZD) strategy on strategy decision, which can be categorized into cooperation and defecting. Our experimental results demonstrate the effectiveness of the proposed strategy decision method.

cs.GT↗
arXiv2018

Image-to-GPS Verification Through A Bottom-Up Pattern Matching Network

The image-to-GPS verification problem asks whether a given image is taken at a claimed GPS location. In this paper, we treat it as an image verification problem -- whether a query image is taken at the same place as a reference image retrieved at the claimed GPS location. We make three major contributions: 1) we propose a novel custom bottom-up pattern matching (BUPM) deep neural network solution; 2) we demonstrate that the verification can be directly done by cross-checking a perspective-looking query image and a panorama reference image, and 3) we collect and clean a dataset of 30K pairs query and reference. Our experimental results show that the proposed BUPM solution outperforms the state-of-the-art solutions in terms of both verification and localization.

cs.CV↗
arXiv2018

GD-1: the relic of an old metal-poor globular cluster

Combining data from Gaia DR2, SDSS DR14 and LAMOST DR6, we update the fit to model of the properties of the stellar stream GD-1 and find that it has an age of $\sim 13$ Gyr, [Fe/H] of $-2.2\pm 0.12$, and a distance from the sun of $\sim 8$ kpc. We tabulate 6D phase-space fiducial points along the GD-1 stream orbit over a 90$^\circ$ arc. The fitted orbit shows that the stream has an eccentricity $e \sim 0.3$, perigalacticon of 14.2 kpc, apogalacticon of 27.0 kpc and inclination $i \sim 40^{\circ}$. There is evidence along the arc for 4 candidate stellar overdensities, one candidate gap, two candidate stellar underdensities and is cut off at $ϕ_1 \sim 2^{\circ}$ (in the stream-aligned $(ϕ_1,ϕ_2)$ coordinate system of \citet{kop10}). The spur originating at $ϕ_1 \sim -40^{\circ}$ implies stars were pulled away from the stream trace by an encounter (potentially a dark matter subhalo). The narrowest place ( FWHM $\sim 44.6$ pc) of the GD-1 trace is at $(ϕ_1, ϕ_2^c) \sim (-14^{\circ}, 0.15^{\circ})$, which is $\sim (178.18^{\circ}, 52.19^{\circ})$ in (R. A., Decl.), where the progenitor is possibly located. We also find six BHB and 10 BS spectroscopic stars in the GD-1 stream.

astro-ph.GA↗
arXiv2018

Bidirectional Conditional Generative Adversarial Networks

Conditional Generative Adversarial Networks (cGANs) are generative models that can produce data samples ($x$) conditioned on both latent variables ($z$) and known auxiliary information ($c$). We propose the Bidirectional cGAN (BiCoGAN), which effectively disentangles $z$ and $c$ in the generation process and provides an encoder that learns inverse mappings from $x$ to both $z$ and $c$, trained jointly with the generator and the discriminator. We present crucial techniques for training BiCoGANs, which involve an extrinsic factor loss along with an associated dynamically-tuned importance weight. As compared to other encoder-based cGANs, BiCoGANs encode $c$ more accurately, and utilize $z$ and $c$ more effectively and in a more disentangled way to generate samples.

cs.LG↗
arXiv2018

Spatial-Temporal Densely Connected Convolutional Networks: An Application to CO2 Leakage Detection

In carbon capture and sequestration, building an effective monitoring method is a crucial step to detect and respond to CO2 leakage. CO2 leakage detection methods rely on geophysical observations and monitoring sensor network. However, traditional methods usually require physical models to be interpreted by experts, and the accuracy of these methods will be restricted by different application conditions. In this paper, we develop a novel data-driven detection method based on densely connected convolutional networks. Our detection method learns a mapping relation between seismic data and the CO2 leakage mass. To account for the spatial and temporal characteristics of seismic data, we design a novel network architecture by combining 1-D and 2-D convolutional neural networks together. To overcome the expensive computational cost, we further apply a densely-connecting policy to our network architecture to reduce the network parameters. We employ our detection method to synthetic seismic datasets using Kimberlina model. The numerical results show that our leakage detection method accurately detects the leakage mass. Therefore, our novel CO2 leakage detection method has great potential for monitoring CO2 storage.

physics.geo-ph↗