PyTorchKR Papers

최신 논문

총 10,163편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2006.02768공식 코드 · Papers with Code

Weight Pruning via Adaptive Sparsity Loss

Pruning neural networks has regained interest in recent years as a means to compress state-of-the-art deep neural networks and enable their deployment on resource-constrained devices. In this paper, we propose a robust compressive learning framework that efficiently prunes network parameters during training with minimal computational overhead. We incorporate fast mechanisms to prune individual layers and build upon these to automatically prune the entire network under a user-defined budget constraint. Key to our end-to-end network pruning approach is the formulation of an intuitive and easy-to-implement adaptive sparsity loss that is used to explicitly control sparsity during training, enabling efficient budget-aware optimization. Extensive experiments demonstrate the effectiveness of the proposed framework for image classification on the CIFAR and ImageNet datasets using different architectures, including AlexNet, ResNets and Wide ResNets.

arXiv:2006.02666

Deep Sequential Feature Learning in Clinical Image Classification of Infectious Keratitis

Infectious keratitis is the most common entities of corneal diseases, in which pathogen grows in the cornea leading to inflammation and destruction of the corneal tissues. Infectious keratitis is a medical emergency, for which a rapid and accurate diagnosis is needed for speedy initiation of prompt and precise treatment to halt the disease progress and to limit the extent of corneal damage; otherwise it may develop sight-threatening and even eye-globe-threatening condition. In this paper, we propose a sequential-level deep learning model to effectively discriminate the distinction and subtlety of infectious corneal disease via the classification of clinical images. In this approach, we devise an appropriate mechanism to preserve the spatial structures of clinical images and disentangle the informative features for clinical image classification of infectious keratitis. In competition with 421 ophthalmologists, the performance of the proposed sequential-level deep model achieved 80.00% diagnostic accuracy, far better than the 49.27% diagnostic accuracy achieved by ophthalmologists over 120 test images.

arXiv:2006.02330

Learning Multi-Modal Nonlinear Embeddings: Performance Bounds and an Algorithm

While many approaches exist in the literature to learn low-dimensional representations for data collections in multiple modalities, the generalizability of multi-modal nonlinear embeddings to previously unseen data is a rather overlooked subject. In this work, we first present a theoretical analysis of learning multi-modal nonlinear embeddings in a supervised setting. Our performance bounds indicate that for successful generalization in multi-modal classification and retrieval problems, the regularity of the interpolation functions extending the embedding to the whole data space is as important as the between-class separation and cross-modal alignment criteria. We then propose a multi-modal nonlinear representation learning algorithm that is motivated by these theoretical findings, where the embeddings of the training samples are optimized jointly with the Lipschitz regularity of the interpolators. Experimental comparison to recent multi-modal and single-modal learning algorithms suggests that the proposed method yields promising performance in multi-modal image classification and cross-modal image-text retrieval applications.

arXiv:2006.02026

Image Classification in the Dark using Quanta Image Sensors

State-of-the-art image classifiers are trained and tested using well-illuminated images. These images are typically captured by CMOS image sensors with at least tens of photons per pixel. However, in dark environments when the photon flux is low, image classification becomes difficult because the measured signal is suppressed by noise. In this paper, we present a new low-light image classification solution using Quanta Image Sensors (QIS). QIS are a new type of image sensors that possess photon counting ability without compromising on pixel size and spatial resolution. Numerous studies over the past decade have demonstrated the feasibility of QIS for low-light imaging, but their usage for image classification has not been studied. This paper fills the gap by presenting a student-teacher learning scheme which allows us to classify the noisy QIS raw data. We show that with student-teacher learning, we are able to achieve image classification at a photon level of one photon per pixel or lower. Experimental results verify the effectiveness of the proposed method compared to existing solutions.

arXiv:2006.01250

RUHSNet: 3D Object Detection Using Lidar Data in Real Time

In this work, we address the problem of 3D object detection from point cloud data in real time. For autonomous vehicles to work, it is very important for the perception component to detect the real world objects with both high accuracy and fast inference. We propose a novel neural network architecture along with the training and optimization details for detecting 3D objects in point cloud data. We compare the results with different backbone architectures including the standard ones like VGG, ResNet, Inception with our backbone. Also we present the optimization and ablation studies including designing an efficient anchor. We use the Kitti 3D Birds Eye View dataset for benchmarking and validating our results. Our work surpasses the state of the art in this domain both in terms of average precision and speed running at > 30 FPS. This makes it a feasible option to be deployed in real time applications including self driving cars.

arXiv:2006.01045

A Hierarchical Deep Convolutional Neural Network and Gated Recurrent Unit Framework for Structural Damage Detection

Structural damage detection has become an interdisciplinary area of interest for various engineering fields, while the available damage detection methods are being in the process of adapting machine learning concepts. Most machine learning based methods heavily depend on extracted ``hand-crafted" features that are manually selected in advance by domain experts and then, fixed. Recently, deep learning has demonstrated remarkable performance on traditional challenging tasks, such as image classification, object detection, etc., due to the powerful feature learning capabilities. This breakthrough has inspired researchers to explore deep learning techniques for structural damage detection problems. However, existing methods have considered either spatial relation (e.g., using convolutional neural network (CNN)) or temporal relation (e.g., using long short term memory network (LSTM)) only. In this work, we propose a novel Hierarchical CNN and Gated recurrent unit (GRU) framework to model both spatial and temporal relations, termed as HCG, for structural damage detection. Specifically, CNN is utilized to model the spatial relations and the short-term temporal dependencies among sensors, while the output features of CNN are fed into the GRU to learn the long-term temporal dependencies jointly. Extensive experiments on IASC-ASCE structural health monitoring benchmark and scale model of three-span continuous rigid frame bridge structure datasets have shown that our proposed HCG outperforms other existing methods for structural damage detection significantly.

arXiv:2006.01043

BadNL: Backdoor Attacks against NLP Models with Semantic-preserving Improvements

Deep neural networks (DNNs) have progressed rapidly during the past decade and have been deployed in various real-world applications. Meanwhile, DNN models have been shown to be vulnerable to security and privacy attacks. One such attack that has attracted a great deal of attention recently is the backdoor attack. Specifically, the adversary poisons the target model's training set to mislead any input with an added secret trigger to a target class. Previous backdoor attacks predominantly focus on computer vision (CV) applications, such as image classification. In this paper, we perform a systematic investigation of backdoor attack on NLP models, and propose BadNL, a general NLP backdoor attack framework including novel attack methods. Specifically, we propose three methods to construct triggers, namely BadChar, BadWord, and BadSentence, including basic and semantic-preserving variants. Our attacks achieve an almost perfect attack success rate with a negligible effect on the original model's utility. For instance, using the BadChar, our backdoor attack achieves a 98.9% attack success rate with yielding a utility improvement of 1.5% on the SST-5 dataset when only poisoning 3% of the original set. Moreover, we conduct a user study to prove that our triggers can well preserve the semantics from humans perspective.

arXiv:2006.00996

Latent Domain Learning with Dynamic Residual Adapters

A practical shortcoming of deep neural networks is their specialization to a single task and domain. While recent techniques in domain adaptation and multi-domain learning enable the learning of more domain-agnostic features, their success relies on the presence of domain labels, typically requiring manual annotation and careful curation of datasets. Here we focus on a less explored, but more realistic case: learning from data from multiple domains, without access to domain annotations. In this scenario, standard model training leads to the overfitting of large domains, while disregarding smaller ones. We address this limitation via dynamic residual adapters, an adaptive gating mechanism that helps account for latent domains, coupled with an augmentation strategy inspired by recent style transfer techniques. Our proposed approach is examined on image classification tasks containing multiple latent domains, and we showcase its ability to obtain robust performance across these. Dynamic residual adapters significantly outperform off-the-shelf networks with much larger capacity, and can be incorporated seamlessly with existing architectures in an end-to-end manner.

arXiv:2006.00234

Integrating global spatial features in CNN based Hyperspectral/SAR imagery classification

The land cover classification has played an important role in remote sensing because it can intelligently identify things in one huge remote sensing image to reduce the work of humans. However, a lot of classification methods are designed based on the pixel feature or limited spatial feature of the remote sensing image, which limits the classification accuracy and universality of their methods. This paper proposed a novel method to take into the information of remote sensing image, i.e., geographic latitude-longitude information. In addition, a dual-branch convolutional neural network (CNN) classification method is designed in combination with the global information to mine the pixel features of the image. Then, the features of the two neural networks are fused with another fully neural network to realize the classification of remote sensing images. Finally, two remote sensing images are used to verify the effectiveness of our method, including hyperspectral imaging (HSI) and polarimetric synthetic aperture radar (PolSAR) imagery. The result of the proposed method is superior to the traditional single-channel convolutional neural network.

arXiv:2005.14707공식 코드 · Papers with Code

Towards Context-Agnostic Learning Using Synthetic Data

We propose a novel setting for learning, where the input domain is the image of a map defined on the product of two sets, one of which completely determines the labels. We derive a new risk bound for this setting that decomposes into a bias and an error term, and exhibits a surprisingly weak dependence on the true labels. Inspired by these results, we present an algorithm aimed at minimizing the bias term by exploiting the ability to sample from each set independently. We apply our setting to visual classification tasks, where our approach enables us to train classifiers on datasets that consist entirely of a single synthetic example of each class. On several standard benchmarks for real-world image classification, we achieve robust performance in the context-agnostic setting, with good generalization to real world domains, whereas training directly on real world data without our techniques yields classifiers that are brittle to perturbations of the background.

arXiv:2005.14506공식 코드 · Papers with Code

Deep convolutional tensor network

Neural networks have achieved state of the art results in many areas, supposedly due to parameter sharing, locality, and depth. Tensor networks (TNs) are linear algebraic representations of quantum many-body states based on their entanglement structure. TNs have found use in machine learning. We devise a novel TN based model called Deep convolutional tensor network (DCTN) for image classification, which has parameter sharing, locality, and depth. It is based on the Entangled plaquette states (EPS) TN. We show how EPS can be implemented as a backpropagatable layer. We test DCTN on MNIST, FashionMNIST, and CIFAR10 datasets. A shallow DCTN performs well on MNIST and FashionMNIST and has a small parameter count. Unfortunately, depth increases overfitting and thus decreases test accuracy. Also, DCTN of any depth performs badly on CIFAR10 due to overfitting. It is to be determined why. We discuss how the hyperparameters of DCTN affect its training and overfitting.

arXiv:2005.14288공식 코드 · Papers with Code

ePillID Dataset: A Low-Shot Fine-Grained Benchmark for Pill Identification

Identifying prescription medications is a frequent task for patients and medical professionals; however, this is an error-prone task as many pills have similar appearances (e.g. white round pills), which increases the risk of medication errors. In this paper, we introduce ePillID, the largest public benchmark on pill image recognition, composed of 13k images representing 9804 appearance classes (two sides for 4902 pill types). For most of the appearance classes, there exists only one reference image, making it a challenging low-shot recognition setting. We present our experimental setup and evaluation results of various baseline models on the benchmark. The best baseline using a multi-head metric-learning approach with bilinear features performed remarkably well; however, our error analysis suggests that they still fail to distinguish particularly confusing classes. The code and data are available at https://github.com/usuyama/ePillID-benchmark.

arXiv:2005.14260

Overview: Computer vision and machine learning for microstructural characterization and analysis

The characterization and analysis of microstructure is the foundation of microstructural science, connecting the materials structure to its composition, process history, and properties. Microstructural quantification traditionally involves a human deciding a priori what to measure and then devising a purpose-built method for doing so. However, recent advances in data science, including computer vision (CV) and machine learning (ML) offer new approaches to extracting information from microstructural images. This overview surveys CV approaches to numerically encode the visual information contained in a microstructural image, which then provides input to supervised or unsupervised ML algorithms that find associations and trends in the high-dimensional image representation. CV/ML systems for microstructural characterization and analysis span the taxonomy of image analysis tasks, including image classification, semantic segmentation, object detection, and instance segmentation. These tools enable new approaches to microstructural analysis, including the development of new, rich visual metrics and the discovery of processing-microstructure-property relationships.

arXiv:2005.14236공식 코드 · Papers with Code

Fuzziness-based Spatial-Spectral Class Discriminant Information Preserving Active Learning for Hyperspectral Image Classification

Traditional Active/Self/Interactive Learning for Hyperspectral Image Classification (HSIC) increases the size of the training set without considering the class scatters and randomness among the existing and new samples. Second, very limited research has been carried out on joint spectral-spatial information and finally, a minor but still worth mentioning is the stopping criteria which not being much considered by the community. Therefore, this work proposes a novel fuzziness-based spatial-spectral within and between for both local and global class discriminant information preserving (FLG) method. We first investigate a spatial prior fuzziness-based misclassified sample information. We then compute the total local and global for both within and between class information and formulate it in a fine-grained manner. Later this information is fed to a discriminative objective function to query the heterogeneous samples which eliminate the randomness among the training samples. Experimental results on benchmark HSI datasets demonstrate the effectiveness of the FLG method on Generative, Extreme Learning Machine and Sparse Multinomial Logistic Regression (SMLR)-LORSAL classifiers.

arXiv:2005.13746공식 코드 · Papers with Code

Tensor decomposition to Compress Convolutional Layers in Deep Learning

Feature extraction for tensor data serves as an important step in many tasks such as anomaly detection, process monitoring, image classification, and quality control. Although many methods have been proposed for tensor feature extraction, there are still two challenges that need to be addressed: 1) how to reduce the computation cost for high dimensional and large volume tensor data; 2) how to interpret the output features and evaluate their significance. {The most recent methods in deep learning, such as Convolutional Neural Network (CNN), have shown outstanding performance in analyzing tensor data, but their wide adoption is still hindered by model complexity and lack of interpretability. To fill this research gap, we propose to use CP-decomposition to approximately compress the convolutional layer (CPAC-Conv layer) in deep learning. The contributions of our work could be summarized into three aspects: (1) we adapt CP-decomposition to compress convolutional kernels and derive the expressions of both forward and backward propagations for our proposed CPAC-Conv layer; (2) compared with the original convolutional layer, the proposed CPAC-Conv layer can reduce the number of parameters without decaying prediction performance. It can combine with other layers to build novel deep Neural Networks; (3) the value of decomposed kernels indicates the significance of the corresponding feature map, which provides us with insights to guide feature selection.

arXiv:2005.13702공식 코드 · Papers with Code

On the Difficulty of Membership Inference Attacks

Recent studies propose membership inference (MI) attacks on deep models, where the goal is to infer if a sample has been used in the training process. Despite their apparent success, these studies only report accuracy, precision, and recall of the positive class (member class). Hence, the performance of these attacks have not been clearly reported on negative class (non-member class). In this paper, we show that the way the MI attack performance has been reported is often misleading because they suffer from high false positive rate or false alarm rate (FAR) that has not been reported. FAR shows how often the attack model mislabel non-training samples (non-member) as training (member) ones. The high FAR makes MI attacks fundamentally impractical, which is particularly more significant for tasks such as membership inference where the majority of samples in reality belong to the negative (non-training) class. Moreover, we show that the current MI attack models can only identify the membership of misclassified samples with mediocre accuracy at best, which only constitute a very small portion of training samples. We analyze several new features that have not been comprehensively explored for membership inference before, including distance to the decision boundary and gradient norms, and conclude that deep models' responses are mostly similar among train and non-train samples. We conduct several experiments on image classification tasks, including MNIST, CIFAR-10, CIFAR-100, and ImageNet, using various model architecture, including LeNet, AlexNet, ResNet, etc. We show that the current state-of-the-art MI attacks cannot achieve high accuracy and low FAR at the same time, even when the attacker is given several advantages. The source code is available at https://github.com/shrezaei/MI-Attack.

arXiv:2005.13630공식 코드 · Papers with Code

Explaining Neural Networks by Decoding Layer Activations

We present a `CLAssifier-DECoder' architecture (\emph{ClaDec}) which facilitates the comprehension of the output of an arbitrary layer in a neural network (NN). It uses a decoder to transform the non-interpretable representation of the given layer to a representation that is more similar to the domain a human is familiar with. In an image recognition problem, one can recognize what information is represented by a layer by contrasting reconstructed images of \emph{ClaDec} with those of a conventional auto-encoder(AE) serving as reference. We also extend \emph{ClaDec} to allow the trade-off between human interpretability and fidelity. We evaluate our approach for image classification using Convolutional NNs. We show that reconstructed visualizations using encodings from a classifier capture more relevant information for classification than conventional AEs. Relevant code is available at \url{https://github.com/JohnTailor/ClaDec}

arXiv:2005.13461

Peri-Net-Pro: The neural processes with quantified uncertainty for crack patterns

This paper uses the peridynamic theory, which is well-suited to crack studies, to predict the crack patterns in a moving disk and classify them according to the modes and finally perform regression analysis. In that way, the crack patterns are obtained according to each mode by Molecular Dynamic (MD) simulation using the peridynamics. Image classification and regression studies are conducted through Convolutional Neural Networks (CNNs) and the neural processes. First, we increased the amount and quality of the data using peridynamics, which can theoretically compensate for the problems of the finite element method (FEM) in generating crack pattern images. Second, we did the case study for the PMB, LPS, and VES models that were obtained using the peridynamic theory. Case studies were performed to classify the images using CNNs and determine the PMB, LBS, and VES models' suitability. Finally, we performed the regression analysis for the images of the crack patterns with neural processes to predict the crack patterns. In the regression problem, by representing the results of the variance according to the epochs, it can be confirmed that the result of the variance is decreased by increasing the epoch numbers through the neural processes. The most critical point of this study is that the neural processes make an accurate prediction even if there are missing or insufficient training data.

arXiv:2005.13110공식 코드 · Papers with Code

Evolutionary NAS with Gene Expression Programming of Cellular Encoding

The renaissance of neural architecture search (NAS) has seen classical methods such as genetic algorithms (GA) and genetic programming (GP) being exploited for convolutional neural network (CNN) architectures. While recent work have achieved promising performance on visual perception tasks, the direct encoding scheme of both GA and GP has functional complexity deficiency and does not scale well on large architectures like CNN. To address this, we present a new generative encoding scheme -- $symbolic\ linear\ generative\ encoding$ (SLGE) -- simple, yet powerful scheme which embeds local graph transformations in chromosomes of linear fixed-length string to develop CNN architectures of variant shapes and sizes via evolutionary process of gene expression programming. In experiments, the effectiveness of SLGE is shown in discovering architectures that improve the performance of the state-of-the-art handcrafted CNN architectures on CIFAR-10 and CIFAR-100 image classification tasks; and achieves a competitive classification error rate with the existing NAS methods using less GPU resources.

arXiv:2005.13420공식 코드 · Papers with Code

Discretize-Optimize vs. Optimize-Discretize for Time-Series Regression and Continuous Normalizing Flows

We compare the discretize-optimize (Disc-Opt) and optimize-discretize (Opt-Disc) approaches for time-series regression and continuous normalizing flows (CNFs) using neural ODEs. Neural ODEs are ordinary differential equations (ODEs) with neural network components. Training a neural ODE is an optimal control problem where the weights are the controls and the hidden features are the states. Every training iteration involves solving an ODE forward and another backward in time, which can require large amounts of computation, time, and memory. Comparing the Opt-Disc and Disc-Opt approaches in image classification tasks, Gholami et al. (2019) suggest that Disc-Opt is preferable due to the guaranteed accuracy of gradients. In this paper, we extend the comparison to neural ODEs for time-series regression and CNFs. Unlike in classification, meaningful models in these tasks must also satisfy additional requirements beyond accurate final-time output, e.g., the invertibility of the CNF. Through our numerical experiments, we demonstrate that with careful numerical treatment, Disc-Opt methods can achieve similar performance as Opt-Disc at inference with drastically reduced training costs. Disc-Opt reduced costs in six out of seven separate problems with training time reduction ranging from 39% to 97%, and in one case, Disc-Opt reduced training from nine days to less than one day.