PyTorchKR Papers

최신 논문

총 12,350편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:1412.4186

An Evaluation of Support Vector Machines as a Pattern Recognition Tool

The purpose of this report is in examining the generalization performance of Support Vector Machines (SVM) as a tool for pattern recognition and object classification. The work is motivated by the growing popularity of the method that is claimed to guarantee a good generalization performance for the task in hand. The method is implemented in MATLAB. SVMs based on various kernels are tested for classifying data from various domains.

arXiv:1412.4175

Optimizing Over Radial Kernels on Compact Manifolds

We tackle the problem of optimizing over all possible positive definite radial kernels on Riemannian manifolds for classification. Kernel methods on Riemannian manifolds have recently become increasingly popular in computer vision. However, the number of known positive definite kernels on manifolds remain very limited. Furthermore, most kernels typically depend on at least one parameter that needs to be tuned for the problem at hand. A poor choice of kernel, or of parameter value, may yield significant performance drop-off. Here, we show that positive definite radial kernels on the unit n-sphere, the Grassmann manifold and Kendall's shape manifold can be expressed in a simple form whose parameters can be automatically optimized within a support vector machine framework. We demonstrate the benefits of our kernel learning algorithm on object, face, action and shape recognition.

arXiv:1412.4174

A Framework for Shape Analysis via Hilbert Space Embedding

We propose a framework for 2D shape analysis using positive definite kernels defined on Kendall's shape manifold. Different representations of 2D shapes are known to generate different nonlinear spaces. Due to the nonlinearity of these spaces, most existing shape classification algorithms resort to nearest neighbor methods and to learning distances on shape spaces. Here, we propose to map shapes on Kendall's shape manifold to a high dimensional Hilbert space where Euclidean geometry applies. To this end, we introduce a kernel on this manifold that permits such a mapping, and prove its positive definiteness. This kernel lets us extend kernel-based algorithms developed for Euclidean spaces, such as SVM, MKL and kernel PCA, to the shape manifold. We demonstrate the benefits of our approach over the state-of-the-art methods on shape classification, clustering and retrieval.

arXiv:1412.3949

CITlab ARGUS for historical handwritten documents

We describe CITlab's recognition system for the HTRtS competition attached to the 14. International Conference on Frontiers in Handwriting Recognition, ICFHR 2014. The task comprises the recognition of historical handwritten documents. The core algorithms of our system are based on multi-dimensional recurrent neural networks (MDRNN) and connectionist temporal classification (CTC). The software modules behind that as well as the basic utility technologies are essentially powered by PLANET's ARGUS framework for intelligent text recognition and image processing.

arXiv:1412.3161

Object-centric Sampling for Fine-grained Image Classification

This paper proposes to go beyond the state-of-the-art deep convolutional neural network (CNN) by incorporating the information from object detection, focusing on dealing with fine-grained image classification. Unfortunately, CNN suffers from over-fiting when it is trained on existing fine-grained image classification benchmarks, which typically only consist of less than a few tens of thousands training images. Therefore, we first construct a large-scale fine-grained car recognition dataset that consists of 333 car classes with more than 150 thousand training images. With this large-scale dataset, we are able to build a strong baseline for CNN with top-1 classification accuracy of 81.6%. One major challenge in fine-grained image classification is that many classes are very similar to each other while having large within-class variation. One contributing factor to the within-class variation is cluttered image background. However, the existing CNN training takes uniform window sampling over the image, acting as blind on the location of the object of interest. In contrast, this paper proposes an \emph{object-centric sampling} (OCS) scheme that samples image windows based on the object location information. The challenge in using the location information lies in how to design powerful object detector and how to handle the imperfectness of detection results. To that end, we design a saliency-aware object detection approach specific for the setting of fine-grained image classification, and the uncertainty of detection results are naturally handled in our OCS scheme. Our framework is demonstrated to be very effective, improving top-1 accuracy to 89.3% (from 81.6%) on the large-scale fine-grained car classification dataset.

arXiv:1412.2106

Consistent optimization of AMS by logistic loss minimization

In this paper, we theoretically justify an approach popular among participants of the Higgs Boson Machine Learning Challenge to optimize approximate median significance (AMS). The approach is based on the following two-stage procedure. First, a real-valued function is learned by minimizing a surrogate loss for binary classification, such as logistic loss, on the training sample. Then, a threshold is tuned on a separate validation sample, by direct optimization of AMS. We show that the regret of the resulting (thresholded) classifier measured with respect to the squared AMS, is upperbounded by the regret of the underlying real-valued function measured with respect to the logistic loss. Hence, we prove that minimizing logistic surrogate is a consistent method of optimizing AMS.

arXiv:1408.6257

Sparse Graph-based Transduction for Image Classification

Motivated by the remarkable successes of Graph-based Transduction (GT) and Sparse Representation (SR), we present a novel Classifier named Sparse Graph-based Classifier (SGC) for image classification. In SGC, SR is leveraged to measure the correlation (similarity) of each two samples and a graph is constructed for encoding these correlations. Then the Laplacian eigenmapping is adopted for deriving the graph Laplacian of the graph. Finally, SGC can be obtained by plugging the graph Laplacian into the conventional GT framework. In the image classification procedure, SGC utilizes the correlations, which are encoded in the learned graph Laplacian, to infer the labels of unlabeled images. SGC inherits the merits of both GT and SR. Compared to SR, SGC improves the robustness and the discriminating power of GT. Compared to GT, SGC sufficiently exploits the whole data. Therefore it alleviates the undercomplete dictionary issue suffered by SR. Four popular image databases are employed for evaluation. The results demonstrate that SGC can achieve a promising performance in comparison with the state-of-the-art classifiers, particularly in the small training sample size case and the noisy sample case.

arXiv:1311.6556

Double Ramp Loss Based Reject Option Classifier

We consider the problem of learning reject option classifiers. The goodness of a reject option classifier is quantified using $0-d-1$ loss function wherein a loss $d \in (0,.5)$ is assigned for rejection. In this paper, we propose {\em double ramp loss} function which gives a continuous upper bound for $(0-d-1)$ loss. Our approach is based on minimizing regularized risk under the double ramp loss using {\em difference of convex (DC) programming}. We show the effectiveness of our approach through experiments on synthetic and benchmark datasets. Our approach performs better than the state of the art reject option classification approaches.

arXiv:1502.05988

Deep Learning for Multi-label Classification

In multi-label classification, the main focus has been to develop ways of learning the underlying dependencies between labels, and to take advantage of this at classification time. Developing better feature-space representations has been predominantly employed to reduce complexity, e.g., by eliminating non-helpful feature attributes from the input space prior to (or during) training. This is an important task, since many multi-label methods typically create many different copies or views of the same input data as they transform it, and considerable memory can be saved by taking advantage of redundancy. In this paper, we show that a proper development of the feature space can make labels less interdependent and easier to model and predict at inference time. For this task we use a deep learning approach with restricted Boltzmann machines. We present a deep network that, in an empirical evaluation, outperforms a number of competitive methods from the literature

arXiv:1412.6257

Gradual training of deep denoising auto encoders

Stacked denoising auto encoders (DAEs) are well known to learn useful deep representations, which can be used to improve supervised training by initializing a deep network. We investigate a training scheme of a deep DAE, where DAE layers are gradually added and keep adapting as additional layers are added. We show that in the regime of mid-sized datasets, this gradual training provides a small but consistent improvement over stacked training in both reconstruction quality and classification error over stacked training on MNIST and CIFAR datasets.

arXiv:1412.6115

Compressing Deep Convolutional Networks using Vector Quantization

Deep convolutional neural networks (CNN) has become the most promising method for object recognition, repeatedly demonstrating record breaking results for image classification and object detection in recent years. However, a very deep CNN generally involves many layers with millions of parameters, making the storage of the network model to be extremely large. This prohibits the usage of deep CNNs on resource limited hardware, especially cell phones or other embedded devices. In this paper, we tackle this model storage issue by investigating information theoretical vector quantization methods for compressing the parameters of CNNs. In particular, we have found in terms of compressing the most storage demanding dense connected layers, vector quantization methods have a clear gain over existing matrix factorization methods. Simply applying k-means clustering to the weights or conducting product quantization can lead to a very good balance between model size and recognition accuracy. For the 1000-category classification task in the ImageNet challenge, we are able to achieve 16-24 times compression of the network with only 1% loss of classification accuracy using the state-of-the-art CNN.

arXiv:1412.6012

CITlab ARGUS for historical data tables

We describe CITlab's recognition system for the ANWRESH-2014 competition attached to the 14. International Conference on Frontiers in Handwriting Recognition, ICFHR 2014. The task comprises word recognition from segmented historical documents. The core components of our system are based on multi-dimensional recurrent neural networks (MDRNN) and connectionist temporal classification (CTC). The software modules behind that as well as the basic utility technologies are essentially powered by PLANET's ARGUS framework for intelligent text recognition and image processing.

arXiv:1412.5710

Multiobjective Optimization of Classifiers by Means of 3-D Convex Hull Based Evolutionary Algorithm

Finding a good classifier is a multiobjective optimization problem with different error rates and the costs to be minimized. The receiver operating characteristic is widely used in the machine learning community to analyze the performance of parametric classifiers or sets of Pareto optimal classifiers. In order to directly compare two sets of classifiers the area (or volume) under the convex hull can be used as a scalar indicator for the performance of a set of classifiers in receiver operating characteristic space. Recently, the convex hull based multiobjective genetic programming algorithm was proposed and successfully applied to maximize the convex hull area for binary classification problems. The contribution of this paper is to extend this algorithm for dealing with higher dimensional problem formulations. In particular, we discuss problems where parsimony (or classifier complexity) is stated as a third objective and multi-class classification with three different true classification rates to be maximized. The design of the algorithm proposed in this paper is inspired by indicator-based evolutionary algorithms, where first a performance indicator for a solution set is established and then a selection operator is designed that complies with the performance indicator. In this case, the performance indicator will be the volume under the convex hull. The algorithm is tested and analyzed in a proof of concept study on different benchmarks that are designed for measuring its capability to capture relevant parts of a convex hull. Further benchmark and application studies on email classification and feature selection round up the analysis and assess robustness and usefulness of the new algorithm in real world settings.

arXiv:1412.5627

Feature extraction from complex networks: A case of study in genomic sequences classification

This work presents a new approach for classification of genomic sequences from measurements of complex networks and information theory. For this, it is considered the nucleotides, dinucleotides and trinucleotides of a genomic sequence. For each of them, the entropy, sum entropy and maximum entropy values are calculated.For each of them is also generated a network, in which the nodes are the nucleotides, dinucleotides or trinucleotides and its edges are estimated by observing the respective adjacency among them in the genomic sequence. In this way, it is generated three networks, for which measures of complex networks are extracted.These measures together with measures of information theory comprise a feature vector representing a genomic sequence. Thus, the feature vector is used for classification by methods such as SVM, MultiLayer Perceptron, J48, IBK, Naive Bayes and Random Forest in order to evaluate the proposed approach.It was adopted coding sequences, intergenic sequences and TSS (Transcriptional Starter Sites) as datasets, for which the better results were obtained by the Random Forest with 91.2%, followed by J48 with 89.1% and SVM with 84.8% of accuracy. These results indicate that the new approach of feature extraction has its value, reaching good levels of classification even considering only the genomic sequences, i.e., no other a priori knowledge about them is considered.

arXiv:1412.4526

Highly Efficient Forward and Backward Propagation of Convolutional Neural Networks for Pixelwise Classification

We present highly efficient algorithms for performing forward and backward propagation of Convolutional Neural Network (CNN) for pixelwise classification on images. For pixelwise classification tasks, such as image segmentation and object detection, surrounding image patches are fed into CNN for predicting the classes of centered pixels via forward propagation and for updating CNN parameters via backward propagation. However, forward and backward propagation was originally designed for whole-image classification. Directly applying it to pixelwise classification in a patch-by-patch scanning manner is extremely inefficient, because surrounding patches of pixels have large overlaps, which lead to a lot of redundant computation. The proposed algorithms eliminate all the redundant computation in convolution and pooling on images by introducing novel d-regularly sparse kernels. It generates exactly the same results as those by patch-by-patch scanning. Convolution and pooling operations with such kernels are able to continuously access memory and can run efficiently on GPUs. A fraction of patches of interest can be chosen from each training image for backward propagation by applying a mask to the error map at the last CNN layer. Its computation complexity is constant with respect to the number of patches sampled from the image. Experiments have shown that our proposed algorithms speed up commonly used patch-by-patch scanning over 1500 times in both forward and backward propagation. The speedup increases with the sizes of images and patches.

arXiv:1412.3506

Road Detection by One-Class Color Classification: Dataset and Experiments

Detecting traversable road areas ahead a moving vehicle is a key process for modern autonomous driving systems. A common approach to road detection consists of exploiting color features to classify pixels as road or background. These algorithms reduce the effect of lighting variations and weather conditions by exploiting the discriminant/invariant properties of different color representations. Furthermore, the lack of labeled datasets has motivated the development of algorithms performing on single images based on the assumption that the bottom part of the image belongs to the road surface. In this paper, we first introduce a dataset of road images taken at different times and in different scenarios using an onboard camera. Then, we devise a simple online algorithm and conduct an exhaustive evaluation of different classifiers and the effect of using different color representation to characterize pixels.

arXiv:1410.0291

A Morphological Analyzer for Japanese Nouns, Verbs and Adjectives

We present an open source morphological analyzer for Japanese nouns, verbs and adjectives. The system builds upon the morphological analyzing capabilities of MeCab to incorporate finer details of classification such as politeness, tense, mood and voice attributes. We implemented our analyzer in the form of a finite state transducer using the open source finite state compiler FOMA toolkit. The source code and tool is available at https://bitbucket.org/skylander/yc-nlplab/.

arXiv:1407.7008

Modeling and Recognition of Smart Grid Faults by a Combined Approach of Dissimilarity Learning and One-Class Classification

Detecting faults in electrical power grids is of paramount importance, either from the electricity operator and consumer viewpoints. Modern electric power grids (smart grids) are equipped with smart sensors that allow to gather real-time information regarding the physical status of all the component elements belonging to the whole infrastructure (e.g., cables and related insulation, transformers, breakers and so on). In real-world smart grid systems, usually, additional information that are related to the operational status of the grid itself are collected such as meteorological information. Designing a suitable recognition (discrimination) model of faults in a real-world smart grid system is hence a challenging task. This follows from the heterogeneity of the information that actually determine a typical fault condition. The second point is that, for synthesizing a recognition model, in practice only the conditions of observed faults are usually meaningful. Therefore, a suitable recognition model should be synthesized by making use of the observed fault conditions only. In this paper, we deal with the problem of modeling and recognizing faults in a real-world smart grid system, which supplies the entire city of Rome, Italy. Recognition of faults is addressed by following a combined approach of multiple dissimilarity measures customization and one-class classification techniques. We provide here an in-depth study related to the available data and to the models synthesized by the proposed one-class classifier. We offer also a comprehensive analysis of the fault recognition results by exploiting a fuzzy set based reliability decision rule.

arXiv:1401.0872

Binary Linear Classification and Feature Selection via Generalized Approximate Message Passing

For the problem of binary linear classification and feature selection, we propose algorithmic approaches to classifier design based on the generalized approximate message passing (GAMP) algorithm, recently proposed in the context of compressive sensing. We are particularly motivated by problems where the number of features greatly exceeds the number of training examples, but where only a few features suffice for accurate classification. We show that sum-product GAMP can be used to (approximately) minimize the classification error rate and max-sum GAMP can be used to minimize a wide variety of regularized loss functions. Furthermore, we describe an expectation-maximization (EM)-based scheme to learn the associated model parameters online, as an alternative to cross-validation, and we show that GAMP's state-evolution framework can be used to accurately predict the misclassification rate. Finally, we present a detailed numerical study to confirm the accuracy, speed, and flexibility afforded by our GAMP-based approaches to binary linear classification and feature selection.

arXiv:1412.8534

Disjunctive Normal Networks

Artificial neural networks are powerful pattern classifiers; however, they have been surpassed in accuracy by methods such as support vector machines and random forests that are also easier to use and faster to train. Backpropagation, which is used to train artificial neural networks, suffers from the herd effect problem which leads to long training times and limit classification accuracy. We use the disjunctive normal form and approximate the boolean conjunction operations with products to construct a novel network architecture. The proposed model can be trained by minimizing an error function and it allows an effective and intuitive initialization which solves the herd-effect problem associated with backpropagation. This leads to state-of-the art classification accuracy and fast training times. In addition, our model can be jointly optimized with convolutional features in an unified structure leading to state-of-the-art results on computer vision problems with fast convergence rates. A GPU implementation of LDNN with optional convolutional features is also available