arXiv subjects

Trung Le

Publications and source records attributed to Trung Le.

138 records · Page 8Linked to original sources

arXiv2019

When Can Neural Networks Learn Connected Decision Regions?

Previous work has questioned the conditions under which the decision regions of a neural network are connected and further showed the implications of the corresponding theory to the problem of adversarial manipulation of classifiers. It has been proven that for a class of activation functions including leaky ReLU, neural networks having a pyramidal structure, that is no layer has more hidden units than the input dimension, produce necessarily connected decision regions. In this paper, we advance this important result by further developing the sufficient and necessary conditions under which the decision regions of a neural network are connected. We then apply our framework to overcome the limits of existing work and further study the capacity to learn connected regions of neural networks for a much wider class of activation functions including those widely used, namely ReLU, sigmoid, tanh, softlus, and exponential linear function.

cs.LG↗

arXiv2017

KGAN: How to Break The Minimax Game in GAN

Generative Adversarial Networks (GANs) were intuitively and attractively explained under the perspective of game theory, wherein two involving parties are a discriminator and a generator. In this game, the task of the discriminator is to discriminate the real and generated (i.e., fake) data, whilst the task of the generator is to generate the fake data that maximally confuses the discriminator. In this paper, we propose a new viewpoint for GANs, which is termed as the minimizing general loss viewpoint. This viewpoint shows a connection between the general loss of a classification problem regarding a convex loss function and a f-divergence between the true and fake data distributions. Mathematically, we proposed a setting for the classification problem of the true and fake data, wherein we can prove that the general loss of this classification problem is exactly the negative f-divergence for a certain convex function f. This allows us to interpret the problem of learning the generator for dismissing the f-divergence between the true and fake data distributions as that of maximizing the general loss which is equivalent to the min-max problem in GAN if the Logistic loss is used in the classification problem. However, this viewpoint strengthens GANs in two ways. First, it allows us to employ any convex loss function for the discriminator. Second, it suggests that rather than limiting ourselves in NN-based discriminators, we can alternatively utilize other powerful families. Bearing this viewpoint, we then propose using the kernel-based family for discriminators. This family has two appealing features: i) a powerful capacity in classifying non-linear nature data and ii) being convex in the feature space. Using the convexity of this family, we can further develop Fenchel duality to equivalently transform the max-min problem to the max-max dual problem.

cs.LG↗

arXiv2017

Multi-Generator Generative Adversarial Nets

We propose a new approach to train the Generative Adversarial Nets (GANs) with a mixture of generators to overcome the mode collapsing problem. The main intuition is to employ multiple generators, instead of using a single one as in the original GAN. The idea is simple, yet proven to be extremely effective at covering diverse data modes, easily overcoming the mode collapse and delivering state-of-the-art results. A minimax formulation is able to establish among a classifier, a discriminator, and a set of generators in a similar spirit with GAN. Generators create samples that are intended to come from the same distribution as the training data, whilst the discriminator determines whether samples are true data or generated by generators, and the classifier specifies which generator a sample comes from. The distinguishing feature is that internal samples are created from multiple generators, and then one of them will be randomly selected as final output similar to the mechanism of a probabilistic mixture model. We term our method Mixture GAN (MGAN). We develop theoretical analysis to prove that, at the equilibrium, the Jensen-Shannon divergence (JSD) between the mixture of generators' distributions and the empirical data distribution is minimal, whilst the JSD among generators' distributions is maximal, hence effectively avoiding the mode collapse. By utilizing parameter sharing, our proposed model adds minimal computational cost to the standard GAN, and thus can also efficiently scale to large-scale datasets. We conduct extensive experiments on synthetic 2D data and natural image databases (CIFAR-10, STL-10 and ImageNet) to demonstrate the superior performance of our MGAN in achieving state-of-the-art Inception scores over latest baselines, generating diverse and appealing recognizable objects at different resolutions, and specializing in capturing different types of objects by generators.

cs.LG↗

arXiv2017

Analogical-based Bayesian Optimization

Some real-world problems revolve to solve the optimization problem \max_{x\in\mathcal{X}}f\left(x\right) where f\left(.\right) is a black-box function and X might be the set of non-vectorial objects (e.g., distributions) where we can only define a symmetric and non-negative similarity score on it. This setting requires a novel view for the standard framework of Bayesian Optimization that generalizes the core insightful spirit of this framework. With this spirit, in this paper, we propose Analogical-based Bayesian Optimization that can maximize black-box function over a domain where only a similarity score can be defined. Our pathway is as follows: we first base on the geometric view of Gaussian Processes (GP) to define the concept of influence level that allows us to analytically represent predictive means and variances of GP posteriors and base on that view to enable replacing kernel similarity by a more genetic similarity score. Furthermore, we also propose two strategies to find a batch of query points that can efficiently handle high dimensional data.

cs.LG↗

arXiv2017

Scalable Support Vector Clustering Using Budget

Owing to its application in solving the difficult and diverse clustering or outlier detection problem, support-based clustering has recently drawn plenty of attention. Support-based clustering method always undergoes two phases: finding the domain of novelty and performing clustering assignment. To find the domain of novelty, the training time given by the current solvers is typically over-quadratic in the training size, and hence precluding the usage of support-based clustering method for large-scale datasets. In this paper, we propose applying Stochastic Gradient Descent (SGD) framework to the first phase of support-based clustering for finding the domain of novelty and a new strategy to perform the clustering assignment. However, the direct application of SGD to the first phase of support-based clustering is vulnerable to the curse of kernelization, that is, the model size linearly grows up with the data size accumulated overtime. To address this issue, we invoke the budget approach which allows us to restrict the model size to a small budget. Our new strategy for clustering assignment enables a fast computation by means of reducing the task of clustering assignment on the full training set to the same task on a significantly smaller set. We also provide a rigorous theoretical analysis about the convergence rate for the proposed method. Finally, we validate our proposed method on the well-known datasets for clustering to show that the proposed method offers a comparable clustering quality while simultaneously achieving significant speedup in comparison with the baselines.

cs.LG↗

arXiv2017

Dual Discriminator Generative Adversarial Nets

We propose in this paper a novel approach to tackle the problem of mode collapse encountered in generative adversarial network (GAN). Our idea is intuitive but proven to be very effective, especially in addressing some key limitations of GAN. In essence, it combines the Kullback-Leibler (KL) and reverse KL divergences into a unified objective function, thus it exploits the complementary statistical properties from these divergences to effectively diversify the estimated density in capturing multi-modes. We term our method dual discriminator generative adversarial nets (D2GAN) which, unlike GAN, has two discriminators; and together with a generator, it also has the analogy of a minimax game, wherein a discriminator rewards high scores for samples from data distribution whilst another discriminator, conversely, favoring data from the generator, and the generator produces data to fool both two discriminators. We develop theoretical analysis to show that, given the maximal discriminators, optimizing the generator of D2GAN reduces to minimizing both KL and reverse KL divergences between data distribution and the distribution induced from the data generated by the generator, hence effectively avoiding the mode collapsing problem. We conduct extensive experiments on synthetic and real-world large-scale datasets (MNIST, CIFAR-10, STL-10, ImageNet), where we have made our best effort to compare our D2GAN with the latest state-of-the-art GAN's variants in comprehensive qualitative and quantitative evaluations. The experimental results demonstrate the competitive and superior performance of our approach in generating good quality and diverse samples over baselines, and the capability of our method to scale up to ImageNet database.

cs.LG↗

arXiv2017

Geometric Enclosing Networks

Training model to generate data has increasingly attracted research attention and become important in modern world applications. We propose in this paper a new geometry-based optimization approach to address this problem. Orthogonal to current state-of-the-art density-based approaches, most notably VAE and GAN, we present a fresh new idea that borrows the principle of minimal enclosing ball to train a generator G\left(\bz\right) in such a way that both training and generated data, after being mapped to the feature space, are enclosed in the same sphere. We develop theory to guarantee that the mapping is bijective so that its inverse from feature space to data space results in expressive nonlinear contours to describe the data manifold, hence ensuring data generated are also lying on the data manifold learned from training data. Our model enjoys a nice geometric interpretation, hence termed Geometric Enclosing Networks (GEN), and possesses some key advantages over its rivals, namely simple and easy-to-control optimization formulation, avoidance of mode collapsing and efficiently learn data manifold representation in a completely unsupervised manner. We conducted extensive experiments on synthesis and real-world datasets to illustrate the behaviors, strength and weakness of our proposed GEN, in particular its ability to handle multi-modal data and quality of generated data.

cs.LG↗

arXiv2017

Approximation Vector Machines for Large-scale Online Learning

One of the most challenging problems in kernel online learning is to bound the model size and to promote the model sparsity. Sparse models not only improve computation and memory usage, but also enhance the generalization capacity, a principle that concurs with the law of parsimony. However, inappropriate sparsity modeling may also significantly degrade the performance. In this paper, we propose Approximation Vector Machine (AVM), a model that can simultaneously encourage the sparsity and safeguard its risk in compromising the performance. When an incoming instance arrives, we approximate this instance by one of its neighbors whose distance to it is less than a predefined threshold. Our key intuition is that since the newly seen instance is expressed by its nearby neighbor the optimal performance can be analytically formulated and maintained. We develop theoretical foundations to support this intuition and further establish an analysis to characterize the gap between the approximation and optimal solutions. This gap crucially depends on the frequency of approximation and the predefined threshold. We perform the convergence analysis for a wide spectrum of loss functions including Hinge, smooth Hinge, and Logistic for classification task, and $l_1$, $l_2$, and $ε$-insensitive for regression task. We conducted extensive experiments for classification task in batch and online modes, and regression task in online mode over several benchmark datasets. The results show that our proposed AVM achieved a comparable predictive performance with current state-of-the-art methods while simultaneously achieving significant computational speed-up due to the ability of the proposed AVM in maintaining the model size.

cs.LG↗

arXiv2017

Scalable Semi-supervised Learning with Graph-based Kernel Machine

Acquiring labels are often costly, whereas unlabeled data are usually easy to obtain in modern machine learning applications. Semi-supervised learning provides a principled machine learning framework to address such situations, and has been applied successfully in many real-word applications and industries. Nonetheless, most of existing semi-supervised learning methods encounter two serious limitations when applied to modern and large-scale datasets: computational burden and memory usage demand. To this end, we present in this paper the Graph-based semi-supervised Kernel Machine (GKM), a method that leverages the generalization ability of kernel-based method with the geometrical and distributive information formulated through a spectral graph induced from data for semi-supervised learning purpose. Our proposed GKM can be solved directly in the primal form using the Stochastic Gradient Descent method with the ideal convergence rate $O(\frac{1}{T})$. Besides, our formulation is suitable for a wide spectrum of important loss functions in the literature of machine learning (e.g., Hinge, smooth Hinge, Logistic, L1, and ε-insensitive) and smoothness functions (i.e., $l_p(t) = |t|^p$ with $p\ge1$). We further show that the well-known Laplacian Support Vector Machine is a special case of our formulation. We validate our proposed method on several benchmark datasets to demonstrate that GKM is appropriate for the large-scale datasets since it is optimal in memory usage and yields superior classification accuracy whilst simultaneously achieving a significant computation speed-up in comparison with the state-of-the-art baselines.

cs.LG↗

arXiv2011

Vortex formation and instability in the left ventricle

We study the formation of the mitral vortex ring during early diastolic filling in a patient-specific left ventricle (LV) using direct numerical simulation. The geometry of the left ventricle is reconstructed from Magnetic Resonance Imaging (MRI) data of a healthy human subject. The left ventricular kinematics is modeled via a cell-based activation methodology, which is inspired by cardiac electro-physiology and yields physiologic LV wall motion. In the fluid dynamics videos, we describe in detail the three-dimensional structure of the mitral vortex ring, which is formed during early diastolic filling. The ring starts to deform as it propagates toward the apex of the heart and becomes inclined. The trailing secondary vortex tubes are formed as the result of interaction between the vortex ring and the LV wall. These vortex tubes wrap around the circumference and begin to interact with and destabilize the mitral vortex ring. At the end of diastole, the vortex ring impinges on the LV wall and the large-scale intraventricular flow rotates in clockwise direction. We show for the first time that the mitral vortex ring evolution is dominated by a number of vortex-vortex and vortex-wall interactions, including lateral straining and deformation of vortex ring, the interaction of two vortex tubes with unequal strengths, helicity polarization of vortex tubes and twisting instabilities of the vortex cores.

physics.flu-dyn↗

arXiv2011

Background Study on nu_e Appearance from a nu_mu Beam in Very Long Baseline Neutrino Oscillation Experiments with a Large Water Cherenkov Detector

There is a growing interest in very long baseline neutrino oscillation experimentation using accelerator produced neutrino beam as a machinery to probe the last three unmeasured neutrino oscillation parameters: the mixing angle theta_13, the possible CP violating phase delta_CP and the mass hierarchy, namely, the sign of delta-m^2_32. Water Cherenkov detectors such as IMB, Kamiokande and Super-Kamiokande have shown to be very successful at detecting neutrino interactions. Scaling up this technology may continue to provide the required performance for the next generation of experiments. This report presents the latest effort to demonstrate that a next generation (> 100 kton) water Cherenkov detector can be used effectively for the rather difficult task of detecting nu_e events from the neutrino oscillation nu_mu -> nu_e despite the large expected potential background resulting from pi^0 events produced via neutral current interactions.

hep-ex↗

arXiv2009

Overview of the T2K long baseline neutrino oscillation experiment

Neutrino oscillations were discovered by atmospheric and solar neutrino experiments, and have been confirmed by experiments using neutrinos from accelerators and nuclear reactors. It has been found that there are large mixing angles in the $ν_e \to ν_μ$ and $ν_μ\to ν_τ$ oscillations. The third mixing angle $θ_{13}$, which parameterizes the mixing between the first and the third generation, is constrainted to be small by the CHOOZ experiment result. The T2K experiment is a long baseline neutrino oscillation experiment that uses intense neutrino beam produced at J-PARC and Super-Kamiokande detector at 295 km as the far detector to measure $θ_{13}$ using $ν_e$ appearance. In this talk, we will give an overview of the experiment.

hep-ex↗