arXiv subjects

Kun He

Publications and source records attributed to Kun He.

At least 181 records · Page 10Linked to original sources

arXiv2020

Dynamic Graph Correlation Learning for Disease Diagnosis with Incomplete Labels

Disease diagnosis on chest X-ray images is a challenging multi-label classification task. Previous works generally classify the diseases independently on the input image without considering any correlation among diseases. However, such correlation actually exists, for example, Pleural Effusion is more likely to appear when Pneumothorax is present. In this work, we propose a Disease Diagnosis Graph Convolutional Network (DD-GCN) that presents a novel view of investigating the inter-dependency among different diseases by using a dynamic learnable adjacency matrix in graph structure to improve the diagnosis accuracy. To learn more natural and reliable correlation relationship, we feed each node with the image-level individual feature map corresponding to each type of disease. To our knowledge, our method is the first to build a graph over the feature maps with a dynamic adjacency matrix for correlation learning. To further deal with a practical issue of incomplete labels, DD-GCN also utilizes an adaptive loss and a curriculum learning strategy to train the model on incomplete labels. Experimental results on two popular chest X-ray (CXR) datasets show that our prediction accuracy outperforms state-of-the-arts, and the learned graph adjacency matrix establishes the correlation representations of different diseases, which is consistent with expert experience. In addition, we apply an ablation study to demonstrate the effectiveness of each component in DD-GCN.

cs.CV↗

arXiv2020

AT-GAN: An Adversarial Generator Model for Non-constrained Adversarial Examples

Despite the rapid development of adversarial machine learning, most adversarial attack and defense researches mainly focus on the perturbation-based adversarial examples, which is constrained by the input images. In comparison with existing works, we propose non-constrained adversarial examples, which are generated entirely from scratch without any constraint on the input. Unlike perturbation-based attacks, or the so-called unrestricted adversarial attack which is still constrained by the input noise, we aim to learn the distribution of adversarial examples to generate non-constrained but semantically meaningful adversarial examples. Following this spirit, we propose a novel attack framework called AT-GAN (Adversarial Transfer on Generative Adversarial Net). Specifically, we first develop a normal GAN model to learn the distribution of benign data, and then transfer the pre-trained GAN model to estimate the distribution of adversarial examples for the target model. In this way, AT-GAN can learn the distribution of adversarial examples that is very close to the distribution of real data. To our knowledge, this is the first work of building an adversarial generator model that could produce adversarial examples directly from any input noise. Extensive experiments and visualizations show that the proposed AT-GAN can very efficiently generate diverse adversarial examples that are more realistic to human perception. In addition, AT-GAN yields higher attack success rates against adversarially trained models under white-box attack setting and exhibits moderate transferability against black-box models.

cs.CV↗

arXiv2020

Nesterov Accelerated Gradient and Scale Invariance for Adversarial Attacks

Deep learning models are vulnerable to adversarial examples crafted by applying human-imperceptible perturbations on benign inputs. However, under the black-box setting, most existing adversaries often have a poor transferability to attack other defense models. In this work, from the perspective of regarding the adversarial example generation as an optimization process, we propose two new methods to improve the transferability of adversarial examples, namely Nesterov Iterative Fast Gradient Sign Method (NI-FGSM) and Scale-Invariant attack Method (SIM). NI-FGSM aims to adapt Nesterov accelerated gradient into the iterative attacks so as to effectively look ahead and improve the transferability of adversarial examples. While SIM is based on our discovery on the scale-invariant property of deep learning models, for which we leverage to optimize the adversarial perturbations over the scale copies of the input images so as to avoid "overfitting" on the white-box model being attacked and generate more transferable adversarial examples. NI-FGSM and SIM can be naturally integrated to build a robust gradient-based attack to generate more transferable adversarial examples against the defense models. Empirical results on ImageNet dataset demonstrate that our attack methods exhibit higher transferability and achieve higher attack success rates than state-of-the-art gradient-based attacks.

cs.LG↗

arXiv2020

Robust Local Features for Improving the Generalization of Adversarial Training

Adversarial training has been demonstrated as one of the most effective methods for training robust models to defend against adversarial examples. However, adversarially trained models often lack adversarially robust generalization on unseen testing data. Recent works show that adversarially trained models are more biased towards global structure features. Instead, in this work, we would like to investigate the relationship between the generalization of adversarial training and the robust local features, as the robust local features generalize well for unseen shape variation. To learn the robust local features, we develop a Random Block Shuffle (RBS) transformation to break up the global structure features on normal adversarial examples. We continue to propose a new approach called Robust Local Features for Adversarial Training (RLFAT), which first learns the robust local features by adversarial training on the RBS-transformed adversarial examples, and then transfers the robust local features into the training of normal adversarial examples. To demonstrate the generality of our argument, we implement RLFAT in currently state-of-the-art adversarial training frameworks. Extensive experiments on STL-10, CIFAR-10 and CIFAR-100 show that RLFAT significantly improves both the adversarially robust generalization and the standard generalization of adversarial training. Additionally, we demonstrate that our models capture more local features of the object on the images, aligning better with human perception.

cs.CV↗

arXiv2020

Stochastic Item Descent Method for Large Scale Equal Circle Packing Problem

Stochastic gradient descent (SGD) is a powerful method for large-scale optimization problems in the area of machine learning, especially for a finite-sum formulation with numerous variables. In recent years, mini-batch SGD gains great success and has become a standard technique for training deep neural networks fed with big amount of data. Inspired by its success in deep learning, we apply the idea of SGD with batch selection of samples to a classic optimization problem in decision version. Given $n$ unit circles, the equal circle packing problem (ECPP) asks whether there exist a feasible packing that could put all the circles inside a circular container without overlapping. Specifically, we propose a stochastic item descent method (SIDM) for ECPP in large scale, which randomly divides the unit circles into batches and runs Broyden-Fletcher-Goldfarb-Shanno (BFGS) algorithm on the corresponding batch function iteratively to speedup the calculation. We also increase the batch size during the batch iterations to gain higher quality solution. Comparing to the current best packing algorithms, SIDM greatly speeds up the calculation of optimization process and guarantees the solution quality for large scale instances with up to 1500 circle items, while the baseline algorithms usually handle about 300 circle items. The results indicate the highly efficiency of SIDM for this classic optimization problem in large scale, and show potential for other large scale classic optimization problems in which gradient descent is used for optimization.

math.OC↗

arXiv2020

Adaptive Large Neighborhood Search for Circle Bin Packing Problem

We address a new variant of packing problem called the circle bin packing problem (CBPP), which is to find a dense packing of circle items to multiple square bins so as to minimize the number of used bins. To this end, we propose an adaptive large neighborhood search (ALNS) algorithm, which uses our Greedy Algorithm with Corner Occupying Action (GACOA) to construct an initial layout. The greedy solution is usually in a local optimum trap, and ALNS enables multiple neighborhood search that depends on the stochastic annealing schedule to avoid getting stuck in local minimum traps. Specifically, ALNS perturbs the current layout to jump out of a local optimum by iteratively reassigns some circles and accepts the new layout with some probability during the search. The acceptance probability is adjusted adaptively using simulated annealing that fine-tunes the search direction in order to reach the global optimum. We benchmark computational results against GACOA in heterogeneous instances. ALNS always outperforms GACOA in improving the objective function, and in several cases, there is a significant reduction on the number of bins used in the packing.

cs.AI↗

arXiv2020

Hierarchical hidden community detection for protein complex prediction

Motivation: Discovering functional modules in protein-protein interaction (PPI) networks by optimization methods remains a longstanding challenge in biology. Traditional algorithms simply consider strong protein complexes that can be found in the original network by optimizing some metrics, which causes obstacles for the discovery of weak and hidden complexes shielded by stronger complexes. Also, protein complexes are not only in different density but also in a large range of scale, making it extremely difficult to be detected. Toward this objective, we propose a hierarchical hidden community approach to predict protein complexes. Results: We propose a method called HirHide (Hierarchical Hidden Community Detection), which can be combined with traditional community detection methods to enable them to discover hierarchical hidden communities. It is the first community detection algorithm that can find a hierarchical structure as well as hidden structure. We compare the performance of three traditional methods with their HirHide versions. Experimental results show that the HirHide methods using traditional methods as the base algorithms achieve better performance, sometimes even significantly outperform the baselines.

q-bio.MN↗

arXiv2019

Adversarially Robust Generalization Just Requires More Unlabeled Data

Neural network robustness has recently been highlighted by the existence of adversarial examples. Many previous works show that the learned networks do not perform well on perturbed test data, and significantly more labeled data is required to achieve adversarially robust generalization. In this paper, we theoretically and empirically show that with just more unlabeled data, we can learn a model with better adversarially robust generalization. The key insight of our results is based on a risk decomposition theorem, in which the expected robust risk is separated into two parts: the stability part which measures the prediction stability in the presence of perturbations, and the accuracy part which evaluates the standard classification accuracy. As the stability part does not depend on any label information, we can optimize this part using unlabeled data. We further prove that for a specific Gaussian mixture problem, adversarially robust generalization can be almost as easy as the standard generalization in supervised learning if a sufficiently large amount of unlabeled data is provided. Inspired by the theoretical findings, we further show that a practical adversarial training algorithm that leverages unlabeled data can improve adversarial robust generalization on MNIST and Cifar-10.

cs.LG↗

arXiv2019

A Learning based Branch and Bound for Maximum Common Subgraph Problems

Branch-and-bound (BnB) algorithms are widely used to solve combinatorial problems, and the performance crucially depends on its branching heuristic.In this work, we consider a typical problem of maximum common subgraph (MCS), and propose a branching heuristic inspired from reinforcement learning with a goal of reaching a tree leaf as early as possible to greatly reduce the search tree size.Extensive experiments show that our method is beneficial and outperforms current best BnB algorithm for the MCS.

cs.LG↗

arXiv2019

Generalized Majorization-Minimization

Non-convex optimization is ubiquitous in machine learning. Majorization-Minimization (MM) is a powerful iterative procedure for optimizing non-convex functions that works by optimizing a sequence of bounds on the function. In MM, the bound at each iteration is required to \emph{touch} the objective function at the optimizer of the previous bound. We show that this touching constraint is unnecessary and overly restrictive. We generalize MM by relaxing this constraint, and propose a new optimization framework, named Generalized Majorization-Minimization (G-MM), that is more flexible. For instance, G-MM can incorporate application-specific biases into the optimization procedure without changing the objective function. We derive G-MM algorithms for several latent variable models and show empirically that they consistently outperform their MM counterparts in optimizing non-convex objectives. In particular, G-MM algorithms appear to be less sensitive to initialization.

cs.CV↗

arXiv2019

Krylov Subspace Approximation for Local Community Detection in Large Networks

Community detection is an important information mining task to uncover modular structures in large networks. For increasingly common large network data sets, global community detection is prohibitively expensive, and attention has shifted to methods that mine local communities, i.e. identifying all latent members of a particular community from a few labeled seed members. To address such semi-supervised mining task, we systematically develop a local spectral subspace-based community detection method, called LOSP. We define a family of local spectral subspaces based on Krylov subspaces, and seek a sparse indicator for the target community via an $\ell_1$ norm minimization over the Krylov subspace. Variants of LOSP depend on type of random walks with different diffusion speeds, type of random walks, dimension of the local spectral subspace and step of diffusions. The effectiveness of the proposed LOSP approach is theoretically analyzed based on Rayleigh quotients, and it is experimentally verified on a wide variety of real-world networks across social, production and biological domains, as well as on an extensive set of synthetic LFR benchmark datasets.

cs.SI↗

arXiv2019

A New Anchor Word Selection Method for the Separable Topic Discovery

Separable Non-negative Matrix Factorization (SNMF) is an important method for topic modeling, where "separable" assumes every topic contains at least one anchor word, defined as a word that has non-zero probability only on that topic. SNMF focuses on the word co-occurrence patterns to reveal topics by two steps: anchor word selection and topic recovery. The quality of the anchor words strongly influences the quality of the extracted topics. Existing anchor word selection algorithm is to greedily find an approximate convex hull in a high-dimensional word co-occurrence space. In this work, we propose a new method for the anchor word selection by associating the word co-occurrence probability with the words similarity and assuming that the most different words on semantic are potential candidates for the anchor words. Therefore, if the similarity of a word-pair is very low, then the two words are very likely to be the anchor words. According to the statistical information of text corpora, we can get the similarity of all word-pairs. We build the word similarity graph where the nodes correspond to words and weights on edges stand for the word-pair similarity. Following this way, we design a greedy method to find a minimum edge-weight anchor clique of a given size in the graph for the anchor word selection. Extensive experiments on real-world corpus demonstrate the effectiveness of the proposed anchor word selection method that outperforms the common convex hull-based methods on the revealed topic quality. Meanwhile, our method is much faster than typical SNMF based method.

cs.IR↗

arXiv2019

Improving the Generalization of Adversarial Training with Domain Adaptation

By injecting adversarial examples into training data, adversarial training is promising for improving the robustness of deep learning models. However, most existing adversarial training approaches are based on a specific type of adversarial attack. It may not provide sufficiently representative samples from the adversarial domain, leading to a weak generalization ability on adversarial examples from other attacks. Moreover, during the adversarial training, adversarial perturbations on inputs are usually crafted by fast single-step adversaries so as to scale to large datasets. This work is mainly focused on the adversarial training yet efficient FGSM adversary. In this scenario, it is difficult to train a model with great generalization due to the lack of representative adversarial samples, aka the samples are unable to accurately reflect the adversarial domain. To alleviate this problem, we propose a novel Adversarial Training with Domain Adaptation (ATDA) method. Our intuition is to regard the adversarial training on FGSM adversary as a domain adaption task with limited number of target domain samples. The main idea is to learn a representation that is semantically meaningful and domain invariant on the clean domain as well as the adversarial domain. Empirical evaluations on Fashion-MNIST, SVHN, CIFAR-10 and CIFAR-100 demonstrate that ATDA can greatly improve the generalization of adversarial training and the smoothness of the learned models, and outperforms state-of-the-art methods on standard benchmark datasets. To show the transfer ability of our method, we also extend ATDA to the adversarial training on iterative attacks such as PGD-Adversial Training (PAT) and the defense performance is improved considerably.

cs.LG↗

arXiv2019

Effective reinforcement learning based local search for the maximum k-plex problem

The maximum k-plex problem is a computationally complex problem, which emerged from graph-theoretic social network studies. This paper presents an effective hybrid local search for solving the maximum k-plex problem that combines the recently proposed breakout local search algorithm with a reinforcement learning strategy. The proposed approach includes distinguishing features such as: a unified neighborhood search based on the swapping operator, a distance-and-quality reward for actions and a new parameter control mechanism based on reinforcement learning. Extensive experiments for the maximum k-plex problem (k = 2, 3, 4, 5) on 80 benchmark instances from the second DIMACS Challenge demonstrate that the proposed approach can match the best-known results from the literature in all but four problem instances. In addition, the proposed algorithm is able to find 32 new best solutions.

cs.NE↗

arXiv2019

Adaptive Wavelet Clustering for Highly Noisy Data

In this paper we make progress on the unsupervised task of mining arbitrarily shaped clusters in highly noisy datasets, which is a task present in many real-world applications. Based on the fundamental work that first applies a wavelet transform to data clustering, we propose an adaptive clustering algorithm, denoted as AdaWave, which exhibits favorable characteristics for clustering. By a self-adaptive thresholding technique, AdaWave is parameter free and can handle data in various situations. It is deterministic, fast in linear time, order-insensitive, shape-insensitive, robust to highly noisy data, and requires no pre-knowledge on data models. Moreover, AdaWave inherits the ability from the wavelet transform to cluster data in different resolutions. We adopt the "grid labeling" data structure to drastically reduce the memory consumption of the wavelet transform so that AdaWave can be used for relatively high dimensional data. Experiments on synthetic as well as natural datasets demonstrate the effectiveness and efficiency of our proposed method.

cs.DB↗

arXiv2018

Multilevel Language and Vision Integration for Text-to-Clip Retrieval

We address the problem of text-based activity retrieval in video. Given a sentence describing an activity, our task is to retrieve matching clips from an untrimmed video. To capture the inherent structures present in both text and video, we introduce a multilevel model that integrates vision and language features earlier and more tightly than prior work. First, we inject text features early on when generating clip proposals, to help eliminate unlikely clips and thus speed up processing and boost performance. Second, to learn a fine-grained similarity metric for retrieval, we use visual features to modulate the processing of query sentences at the word level in a recurrent neural network. A multi-task loss is also employed by adding query re-generation as an auxiliary task. Our approach significantly outperforms prior work on two challenging benchmarks: Charades-STA and ActivityNet Captions.

cs.CV↗

arXiv2018

Tight bounds for popping algorithms

We sharpen run-time analysis for algorithms under the partial rejection sampling framework. Our method yields improved bounds for: the cluster-popping algorithm for approximating all-terminal network reliability; the cycle-popping algorithm for sampling rooted spanning trees; the sink-popping algorithm for sampling sink-free orientations. In all three applications, our bounds are not only tight in order, but also optimal in constants.

cs.DS↗

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↗