Searcharxiv⌕ Search

arXiv subjects

Liang Zhao

Publications and source records attributed to Liang Zhao.

At least 451 records · Page 25Linked to original sources
arXiv2018

Interpreting Adversarial Robustness: A View from Decision Surface in Input Space

One popular hypothesis of neural network generalization is that the flat local minima of loss surface in parameter space leads to good generalization. However, we demonstrate that loss surface in parameter space has no obvious relationship with generalization, especially under adversarial settings. Through visualizing decision surfaces in both parameter space and input space, we instead show that the geometry property of decision surface in input space correlates well with the adversarial robustness. We then propose an adversarial robustness indicator, which can evaluate a neural network's intrinsic robustness property without testing its accuracy under adversarial attacks. Guided by it, we further propose our robust training method. Without involving adversarial training, our method could enhance network's intrinsic adversarial robustness against various adversarial attacks.

cs.LG↗
arXiv2018

Superfast Low-Rank Approximation and Least Squares Regression

Low Rank Approximation is among most fundamental subjects of numerical linear algebra having important applications to various areas of modern computing and %they range from machine learning theory and %neural networks to data mining and analysis. The known algorithms compute such approximations by using more flops than the input matrix has entries, but we prove that much fewer flops than entries are sufficient in the case of the average input ("flop" stands for "floating point arithmetic operation"). We prove this twice -- for the solutions by means of two distinct algorithms, and we analyze them by applying two different approaches. Our analysis of both algorithms is quite involved, but we devise them mostly by simplifying, combining, and ameliorating the known techniques, although we propose some technical novelties for further enhancing the performance of the popular Cross-Approximation Algorithms. They are highly efficient empirically, and we prove that they are efficient for the average input. We specify some narrow classes of hard inputs for which the presented algorithms fail with high probability even when we randomize them, but we narrow such classes further by means of preprocessing with new sparse and structured multipliers. The average complexity estimates do not cover many realistic input classes, but our formal analysis is in good accordance with the results of our tests applied to benchmark inputs from discretized PDEs and Integral quations and to random inputs. Our work should already be of practical value but also leads to research challenges. At the end we list some of them, propose two novel extensions of our progress -- to the acceleration of the Fast Multipole Method and Conjugate Gradient algorithms, and explore and slightly extend the recent techniques of Osinsky, which enhance the output accuracy of CUR Approximation.

math.NA↗
arXiv2018

A Generic Framework for Interesting Subspace Cluster Detection in Multi-attributed Networks

Detection of interesting (e.g., coherent or anomalous) clusters has been studied extensively on plain or univariate networks, with various applications. Recently, algorithms have been extended to networks with multiple attributes for each node in the real-world. In a multi-attributed network, often, a cluster of nodes is only interesting for a subset (subspace) of attributes, and this type of clusters is called subspace clusters. However, in the current literature, few methods are capable of detecting subspace clusters, which involves concurrent feature selection and network cluster detection. These relevant methods are mostly heuristic-driven and customized for specific application scenarios. In this work, we present a generic and theoretical framework for detection of interesting subspace clusters in large multi-attributed networks. Specifically, we propose a subspace graph-structured matching pursuit algorithm, namely, SG-Pursuit, to address a broad class of such problems for different score functions (e.g., coherence or anomalous functions) and topology constraints (e.g., connected subgraphs and dense subgraphs). We prove that our algorithm 1) runs in nearly-linear time on the network size and the total number of attributes and 2) enjoys rigorous guarantees (geometrical convergence rate and tight error bound) analogous to those of the state-of-the-art algorithms for sparse feature selection problems and subgraph detection problems. As a case study, we specialize SG-Pursuit to optimize a number of well-known score functions for two typical tasks, including detection of coherent dense and anomalous connected subspace clusters in real-world networks. Empirical evidence demonstrates that our proposed generic algorithm SG-Pursuit performs superior over state-of-the-art methods that are designed specifically for these two tasks.

cs.LG↗
arXiv2018

A global weak solution to the Lorentzian harmonic map flow

We investigate a parabolic-elliptic system which is related to a harmonic map from a compact Riemann surface with a smooth boundary into a Lorentzian manifold with a warped product metric. We prove that there exists a unique global weak solution for this system which is regular except for at most finitely many singular points.

math.DG↗
arXiv2018

Linear SLAM: Linearising the SLAM Problems using Submap Joining

The main contribution of this paper is a new submap joining based approach for solving large-scale Simultaneous Localization and Mapping (SLAM) problems. Each local submap is independently built using the local information through solving a small-scale SLAM; the joining of submaps mainly involves solving linear least squares and performing nonlinear coordinate transformations. Through approximating the local submap information as the state estimate and its corresponding information matrix, judiciously selecting the submap coordinate frames, and approximating the joining of a large number of submaps by joining only two maps at a time, either sequentially or in a more efficient Divide and Conquer manner, the nonlinear optimization process involved in most of the existing submap joining approaches is avoided. Thus the proposed submap joining algorithm does not require initial guess or iterations since linear least squares problems have closed-form solutions. The proposed Linear SLAM technique is applicable to feature-based SLAM, pose graph SLAM and D-SLAM, in both two and three dimensions, and does not require any assumption on the character of the covariance matrices. Simulations and experiments are performed to evaluate the proposed Linear SLAM algorithm. Results using publicly available datasets in 2D and 3D show that Linear SLAM produces results that are very close to the best solutions that can be obtained using full nonlinear optimization algorithm started from an accurate initial guess. The C/C++ and MATLAB source codes of Linear SLAM are available on OpenSLAM.

cs.RO↗
arXiv2018

Source-Critical Reinforcement Learning for Transferring Spoken Language Understanding to a New Language

To deploy a spoken language understanding (SLU) model to a new language, language transferring is desired to avoid the trouble of acquiring and labeling a new big SLU corpus. Translating the original SLU corpus into the target language is an attractive strategy. However, SLU corpora consist of plenty of semantic labels (slots), which general-purpose translators cannot handle well, not to mention additional culture differences. This paper focuses on the language transferring task given a tiny in-domain parallel SLU corpus. The in-domain parallel corpus can be used as the first adaptation on the general translator. But more importantly, we show how to use reinforcement learning (RL) to further finetune the adapted translator, where translated sentences with more proper slot tags receive higher rewards. We evaluate our approach on Chinese to English language transferring for SLU systems. The experimental results show that the generated English SLU corpus via adaptation and reinforcement learning gives us over 97% in the slot F1 score and over 84% accuracy in domain classification. It demonstrates the effectiveness of the proposed language transferring method. Compared with naive translation, our proposed method improves domain classification accuracy by relatively 22%, and the slot filling F1 score by relatively more than 71%.

cs.CL↗
arXiv2018

An Improved Speedup Factor for Sporadic Tasks with Constrained Deadlines under Dynamic Priority Scheduling

Schedulability is a fundamental problem in real-time scheduling, but it has to be approximated due to the intrinsic computational hardness. As the most popular algorithm for deciding schedulability on multiprocess platforms, the speedup factor of partitioned-EDF is challenging to analyze and is far from been determined. Partitioned-EDF was first proposed in 2005 by Barush and Fisher [1], and was shown to have a speedup factor at most 3-1/m, meaning that if the input of sporadic tasks is feasible on m processors with speed one, partitioned-EDF will always return succeeded on m processors with speed 3-1/m. In 2011, this upper bound was improved to 2.6322-1/m by Chen and Chakraborty [2], and no more improvements have appeared ever since then. In this paper, we develop a novel method to discretize and regularize sporadic tasks, which enables us to improve, in the case of constrained deadlines, the speedup factor of partitioned-EDF to 2.5556-1/m, very close to the asymptotic lower bound 2.5 in [2].

cs.DS↗
arXiv2018

Parallax Bundle Adjustment on Manifold with Convexified Initialization

Bundle adjustment (BA) with parallax angle based feature parameterization has been shown to have superior performance over BA using inverse depth or XYZ feature forms. In this paper, we propose an improved version of the parallax BA algorithm (PMBA) by extending it to the manifold domain along with observation-ray based objective function. With this modification, the problem formulation faithfully mimics the projective nature in a camera's image formation, BA is able to achieve better convergence, accuracy and robustness. This is particularly useful in handling diverse outdoor environments and collinear motion modes. Capitalizing on these properties, we further propose a pose-graph simplification to PMBA, with significant dimensionality reduction. This pose-graph model is convex in nature, easy to solve and its solution can serve as a good initial guess to the original BA problem which is intrinsically non-convex. We provide theoretical proof that our global initialization strategy can guarantee a near-optimal solution. Using a series of experiments involving diverse environmental conditions and motions, we demonstrate PMBA's superior convergence performance in comparison to other BA methods. We also show that, without incremental initialization or via third-party information, our global initialization process helps to bootstrap the full BA successfully in various scenarios, sequential or out-of-order, including some datasets from the "Bundle Adjustment in the Large" database.

cs.RO↗
arXiv2018

Distributed Self-Paced Learning in Alternating Direction Method of Multipliers

Self-paced learning (SPL) mimics the cognitive process of humans, who generally learn from easy samples to hard ones. One key issue in SPL is the training process required for each instance weight depends on the other samples and thus cannot easily be run in a distributed manner in a large-scale dataset. In this paper, we reformulate the self-paced learning problem into a distributed setting and propose a novel Distributed Self-Paced Learning method (DSPL) to handle large-scale datasets. Specifically, both the model and instance weights can be optimized in parallel for each batch based on a consensus alternating direction method of multipliers. We also prove the convergence of our algorithm under mild conditions. Extensive experiments on both synthetic and real datasets demonstrate that our approach is superior to those of existing methods.

cs.LG↗
arXiv2018

Front Velocity and Front Location of Lock-exchange Gravity Currents Descending a Slope in a Linearly Stratified Environment

Gravity currents descending a slope in a linearly stratified environment can be frequently encountered in nature. However, few studies have quantitatively investigated the evolution process of lock-exchange gravity currents in such environments. A new set of analytical formulae is proposed by integrating both mass conservation and linear momentum equations to determine the front velocity and the front location of a downslope current. Based on the thermal theory, the formula considers the influence of ambient stratification by introducing a newly defined stratification coefficient in the acceleration stage. As for the deceleration stage, the formula is derived by adding a parameter that takes into account the density distribution of the ambient water. The transition point between the acceleration and deceleration stages and the maximum front velocity are also determined by the proposed formulae. Lock-exchange gravity current experiments are conducted in the flume with linear stratifications to provide data for the validation of the formulae. The comparisons between the calculated and measured data in terms of front location and front velocity show satisfactory agreements, which reveal that front velocity presents a rapid acceleration stage and then a deceleration stage in a linearly stratified ambient.

physics.flu-dyn↗
arXiv2018

Deep Graph Translation

Inspired by the tremendous success of deep generative models on generating continuous data like image and audio, in the most recent year, few deep graph generative models have been proposed to generate discrete data such as graphs. They are typically unconditioned generative models which has no control on modes of the graphs being generated. Differently, in this paper, we are interested in a new problem named \emph{Deep Graph Translation}: given an input graph, we want to infer a target graph based on their underlying (both global and local) translation mapping. Graph translation could be highly desirable in many applications such as disaster management and rare event forecasting, where the rare and abnormal graph patterns (e.g., traffic congestions and terrorism events) will be inferred prior to their occurrence even without historical data on the abnormal patterns for this graph (e.g., a road network or human contact network). To achieve this, we propose a novel Graph-Translation-Generative Adversarial Networks (GT-GAN) which will generate a graph translator from input to target graphs. GT-GAN consists of a graph translator where we propose new graph convolution and deconvolution layers to learn the global and local translation mapping. A new conditional graph discriminator has also been proposed to classify target graphs by conditioning on input graphs. Extensive experiments on multiple synthetic and real-world datasets demonstrate the effectiveness and scalability of the proposed GT-GAN.

cs.LG↗
arXiv2018

On the Universal Approximation Property and Equivalence of Stochastic Computing-based Neural Networks and Binary Neural Networks

Large-scale deep neural networks are both memory intensive and computation-intensive, thereby posing stringent requirements on the computing platforms. Hardware accelerations of deep neural networks have been extensively investigated in both industry and academia. Specific forms of binary neural networks (BNNs) and stochastic computing based neural networks (SCNNs) are particularly appealing to hardware implementations since they can be implemented almost entirely with binary operations. Despite the obvious advantages in hardware implementation, these approximate computing techniques are questioned by researchers in terms of accuracy and universal applicability. Also it is important to understand the relative pros and cons of SCNNs and BNNs in theory and in actual hardware implementations. In order to address these concerns, in this paper we prove that the "ideal" SCNNs and BNNs satisfy the universal approximation property with probability 1 (due to the stochastic behavior). The proof is conducted by first proving the property for SCNNs from the strong law of large numbers, and then using SCNNs as a "bridge" to prove for BNNs. Based on the universal approximation property, we further prove that SCNNs and BNNs exhibit the same energy complexity. In other words, they have the same asymptotic energy consumption with the growing of network size. We also provide a detailed analysis of the pros and cons of SCNNs and BNNs for hardware implementations and conclude that SCNNs are more suitable for hardware.

cs.LG↗
arXiv2018

Occlusion Aware Unsupervised Learning of Optical Flow

It has been recently shown that a convolutional neural network can learn optical flow estimation with unsupervised learning. However, the performance of the unsupervised methods still has a relatively large gap compared to its supervised counterpart. Occlusion and large motion are some of the major factors that limit the current unsupervised learning of optical flow methods. In this work we introduce a new method which models occlusion explicitly and a new warping way that facilitates the learning of large motion. Our method shows promising results on Flying Chairs, MPI-Sintel and KITTI benchmark datasets. Especially on KITTI dataset where abundant unlabeled samples exist, our unsupervised method outperforms its counterpart trained with supervised learning.

cs.CV↗
arXiv2018

MIS-SLAM: Real-time Large Scale Dense Deformable SLAM System in Minimal Invasive Surgery Based on Heterogeneous Computing

Real-time simultaneously localization and dense mapping is very helpful for providing Virtual Reality and Augmented Reality for surgeons or even surgical robots. In this paper, we propose MIS-SLAM: a complete real-time large scale dense deformable SLAM system with stereoscope in Minimal Invasive Surgery based on heterogeneous computing by making full use of CPU and GPU. Idled CPU is used to perform ORB- SLAM for providing robust global pose. Strategies are taken to integrate modules from CPU and GPU. We solved the key problem raised in previous work, that is, fast movement of scope and blurry images make the scope tracking fail. Benefiting from improved localization, MIS-SLAM can achieve large scale scope localizing and dense mapping in real-time. It transforms and deforms current model and incrementally fuses new observation while keeping vivid texture. In-vivo experiments conducted on publicly available datasets presented in the form of videos demonstrate the feasibility and practicality of MIS-SLAM for potential clinical purpose.

cs.CV↗
arXiv2018

Learning to Explore with Meta-Policy Gradient

The performance of off-policy learning, including deep Q-learning and deep deterministic policy gradient (DDPG), critically depends on the choice of the exploration policy. Existing exploration methods are mostly based on adding noise to the on-going actor policy and can only explore \emph{local} regions close to what the actor policy dictates. In this work, we develop a simple meta-policy gradient algorithm that allows us to adaptively learn the exploration policy in DDPG. Our algorithm allows us to train flexible exploration behaviors that are independent of the actor policy, yielding a \emph{global exploration} that significantly speeds up the learning process. With an extensive study, we show that our method significantly improves the sample-efficiency of DDPG on a variety of reinforcement learning tasks.

cs.LG↗
arXiv2018

Superfast CUR Matrix Algorithms, Their Pre-Processing and Extensions

We study superfast algorithms that computes low rank approximation of a matrix (hereafter referred to as LRA) that use much fewer memory cells and arithmetic operations than the input matrix has entries. We first specify a family of 2mn matrices of size m*n such that for almost 50% of them any superfast LRA algorithm fails to improve the poor trivial approximation by the matrix filled with zeros, but then we prove that the class of all such hard inputs is narrow - the cross-approximation (hereafter {C-A}) superfast iterations as well as some more primitive superfast algorithms compute reasonably accurate LRAs in their transparent CUR form (i) to any matrix allowing close LRA except for small norm perturbations of matrices of an algebraic variety of a smaller dimension, (ii) to the average matrix allowing close LRA, (iii) to the average sparse matrix allowing close LRA and (iv) with a high probability to any matrix allowing close LRA if it is pre-processed fast with a random Gaussian, SRHT or SRFT multiplier. Moreover empirically the output LRAs remain accurate when we perform the computations superfast by replacing such a multiplier with one of our sparse and structured multipliers. Our techniques, auxiliary results and extensions may be of some independent interest. We analyze C-A and other superfast algorithms twice -- based on two well-known sufficient criteria for obtaining accurate LRAs. We provide a distinct proof in the case of superfast variant of randomized algorithms of [DMM08], improve a decade-old estimate for the norm of the inverse of a Gaussian matrix, prove such an estimate also in the case of a sparse Gaussian matrix, present some novel advanced pre-processing techniques for fast and superfast computation of LRA, and extend our results to dramatic acceleration of the Fast Multipole Method (FMM) and the Conjugate Gradient algorithms.

math.NA↗
arXiv2018

Network Unfolding Map by Edge Dynamics Modeling

The emergence of collective dynamics in neural networks is a mechanism of the animal and human brain for information processing. In this paper, we develop a computational technique using distributed processing elements in a complex network, which are called particles, to solve semi-supervised learning problems. Three actions govern the particles' dynamics: generation, walking, and absorption. Labeled vertices generate new particles that compete against rival particles for edge domination. Active particles randomly walk in the network until they are absorbed by either a rival vertex or an edge currently dominated by rival particles. The result from the model evolution consists of sets of edges arranged by the label dominance. Each set tends to form a connected subnetwork to represent a data class. Although the intrinsic dynamics of the model is a stochastic one, we prove there exists a deterministic version with largely reduced computational complexity; specifically, with linear growth. Furthermore, the edge domination process corresponds to an unfolding map in such way that edges "stretch" and "shrink" according to the vertex-edge dynamics. Consequently, the unfolding effect summarizes the relevant relationships between vertices and the uncovered data classes. The proposed model captures important details of connectivity patterns over the vertex-edge dynamics evolution, in contrast to previous approaches which focused on only vertex or only edge dynamics. Computer simulations reveal that the new model can identify nonlinear features in both real and artificial data, including boundaries between distinct classes and overlapping structures of data.

cs.AI↗