Deep learning has enjoyed a great deal of success because of its ability to learn useful features for tasks such as classification. But there has been less exploration in learning the factors of variation apart from the classification signal. By augmenting autoencoders with simple regularization terms during training, we demonstrate that standard deep architectures can discover and explicitly represent factors of variation beyond those relevant for categorization. We introduce a cross-covariance penalty (XCov) as a method to disentangle factors like handwriting style for digits and subject identity in faces. We demonstrate this on the MNIST handwritten digit database, the Toronto Faces Database (TFD) and the Multi-PIE dataset by generating manipulated instances of the data. Furthermore, we demonstrate these deep networks can extrapolate `hidden' variation in the supervised signal.
Deep convolutional networks have proven to be very successful in learning task specific features that allow for unprecedented performance on various computer vision tasks. Training of such networks follows mostly the supervised learning paradigm, where sufficiently many input-output pairs are required for training. Acquisition of large training sets is one of the key challenges, when approaching a new task. In this paper, we aim for generic feature learning and present an approach for training a convolutional network using only unlabeled data. To this end, we train the network to discriminate between a set of surrogate classes. Each surrogate class is formed by applying a variety of transformations to a randomly sampled 'seed' image patch. In contrast to supervised network training, the resulting feature representation is not class specific. It rather provides robustness to the transformations that have been applied during training. This generic feature representation allows for classification results that outperform the state of the art for unsupervised learning on several popular datasets (STL-10, CIFAR-10, Caltech-101, Caltech-256). While such generic features cannot compete with class specific features from supervised training on a classification task, we show that they are advantageous on geometric matching problems, where they also outperform the SIFT descriptor.
Syntactic features play an essential role in identifying relationship in a sentence. Previous neural network models often suffer from irrelevant information introduced when subjects and objects are in a long distance. In this paper, we propose to learn more robust relation representations from the shortest dependency path through a convolution neural network. We further propose a straightforward negative sampling strategy to improve the assignment of subjects and objects. Experimental results show that our method outperforms the state-of-the-art methods on the SemEval-2010 Task 8 dataset.
Replicated Softmax model, a well-known undirected topic model, is powerful in extracting semantic representations of documents. Traditional learning strategies such as Contrastive Divergence are very inefficient. This paper provides a novel estimator to speed up the learning based on Noise Contrastive Estimate, extended for documents of variant lengths and weighted inputs. Experiments on two benchmarks show that the new estimator achieves great learning efficiency and high accuracy on document retrieval and classification.
Recognizing emotion from speech has become one the active research themes in speech processing and in applications based on human-computer interaction. This paper conducts an experimental study on recognizing emotions from human speech. The emotions considered for the experiments include neutral, anger, joy and sadness. The distinuishability of emotional features in speech were studied first followed by emotion classification performed on a custom dataset. The classification was performed for different classifiers. One of the main feature attribute considered in the prepared dataset was the peak-to-peak distance obtained from the graphical representation of the speech signals. After performing the classification tests on a dataset formed from 30 different subjects, it was found that for getting better accuracy, one should consider the data collected from one person rather than considering the data from a group of people.
In this paper, we proposed a pose estimation system based on rendered image training set, which predicts the pose of objects in real image, with knowledge of object category and tight bounding box. We developed a patch-based multi-class classification algorithm, and an iterative approach to improve the accuracy. We achieved state-of-the-art performance on pose estimation task.
We propose a novel algorithm for optimizing multivariate linear threshold functions as split functions of decision trees to create improved Random Forest classifiers. Standard tree induction methods resort to sampling and exhaustive search to find good univariate split functions. In contrast, our method computes a linear combination of the features at each node, and optimizes the parameters of the linear combination (oblique) split functions by adopting a variant of latent variable SVM formulation. We develop a convex-concave upper bound on the classification loss for a one-level decision tree, and optimize the bound by stochastic gradient descent at each internal node of the tree. Forests of up to 1000 Continuously Optimized Oblique (CO2) decision trees are created, which significantly outperform Random Forest with univariate splits and previous techniques for constructing oblique trees. Experimental results are reported on multi-class classification benchmarks and on Labeled Faces in the Wild (LFW) dataset.
Convolutional Neural Networks demonstrate high performance on ImageNet Large-Scale Visual Recognition Challenges contest. Nevertheless, the published results only show the overall performance for all image classes. There is no further analysis why certain images get worse results and how they could be improved. In this paper, we provide deep performance analysis based on different types of images and point out the weaknesses of convolutional neural networks through experiment. We design a novel multiple paths convolutional neural network, which feeds different versions of images into separated paths to learn more comprehensive features. This model has better presentation for image than the traditional single path model. We acquire better classification results on complex validation set on both top 1 and top 5 scores than the best ILSVRC 2013 classification model.
We present a novel learning method for word embeddings designed for relation classification. Our word embeddings are trained by predicting words between noun pairs using lexical relation-specific features on a large unlabeled corpus. This allows us to explicitly incorporate relation-specific information into the word embeddings. The learned word embeddings are then used to construct feature vectors for a relation classification model. On a well-established semantic relation classification task, our method significantly outperforms a baseline based on a previously introduced word embedding method, and compares favorably to previous state-of-the-art models that use syntactic information or manually constructed external resources.
Palmprint is one of the most useful physiological biometrics that can be used as a powerful means in personal recognition systems. The major features of the palmprints are palm lines, wrinkles and ridges, and many approaches use them in different ways towards solving the palmprint recognition problem. Here we have proposed to use a set of statistical and wavelet-based features; statistical to capture the general characteristics of palmprints; and wavelet-based to find those information not evident in the spatial domain. Also we use two different classification approaches, minimum distance classifier scheme and weighted majority voting algorithm, to perform palmprint matching. The proposed method is tested on a well-known palmprint dataset of 6000 samples and has shown an impressive accuracy rate of 99.65\%-100\% for most scenarios.
Latest results indicate that features learned via convolutional neural networks outperform previous descriptors on classification tasks by a large margin. It has been shown that these networks still work well when they are applied to datasets or recognition tasks different from those they were trained on. However, descriptors like SIFT are not only used in recognition but also for many correspondence problems that rely on descriptor matching. In this paper we compare features from various layers of convolutional neural nets to standard SIFT descriptors. We consider a network that was trained on ImageNet and another one that was trained without supervision. Surprisingly, convolutional neural networks clearly outperform SIFT on descriptor matching. This paper has been merged with arXiv:1406.6909
A method for dimension reduction with clustering, classification, or discriminant analysis is introduced. This mixture model-based approach is based on fitting generalized hyperbolic mixtures on a reduced subspace within the paradigm of model-based clustering, classification, or discriminant analysis. A reduced subspace of the data is derived by considering the extent to which group means and group covariances vary. The members of the subspace arise through linear combinations of the original data, and are ordered by importance via the associated eigenvalues. The observations can be projected onto the subspace, resulting in a set of variables that captures most of the clustering information available. The use of generalized hyperbolic mixtures gives a robust framework capable of dealing with skewed clusters. Although dimension reduction is increasingly in demand across many application areas, the authors are most familiar with biological applications and so two of the five real data examples are within that sphere. Simulated data are also used for illustration. The approach introduced herein can be considered the most general such approach available, and so we compare results to three special and limiting cases. Comparisons with several well established techniques illustrate its promising performance.
We introduce, test and discuss a method for classifying and clustering data modeled as directed graphs. The idea is to start diffusion processes from any subset of a data collection, generating corresponding distributions for reaching points in the network. These distributions take the form of high-dimensional numerical vectors and capture essential topological properties of the original dataset. We show how these diffusion vectors can be successfully applied for getting state-of-the-art accuracies in the problem of extracting pathways from metabolic networks. We also provide a guideline to illustrate how to use our method for classification problems, and discuss important details of its implementation. In particular, we present a simple dimensionality reduction technique that lowers the computational cost of classifying diffusion vectors, while leaving the predictive power of the classification process substantially unaltered. Although the method has very few parameters, the results we obtain show its flexibility and power. This should make it helpful in many other contexts.
We introduce general scattering transforms as mathematical models of deep neural networks with l2 pooling. Scattering networks iteratively apply complex valued unitary operators, and the pooling is performed by a complex modulus. An expected scattering defines a contractive representation of a high-dimensional probability distribution, which preserves its mean-square norm. We show that unsupervised learning can be casted as an optimization of the space contraction to preserve the volume occupied by unlabeled examples, at each layer of the network. Supervised learning and classification are performed with an averaged scattering, which provides scattering estimations for multiple classes.
For polarimetric SAR (PolSAR) image classification, it is a challenge to classify the aggregated terrain types, such as the urban area, into semantic homogenous regions due to sharp bright-dark variations in intensity. The aggregated terrain type is formulated by the similar ground objects aggregated together. In this paper, a polarimetric hierarchical semantic model (PHSM) is firstly proposed to overcome this disadvantage based on the constructions of a primal-level and a middle-level semantic. The primal-level semantic is a polarimetric sketch map which consists of sketch segments as the sparse representation of a PolSAR image. The middle-level semantic is a region map which can extract semantic homogenous regions from the sketch map by exploiting the topological structure of sketch segments. Mapping the region map to the PolSAR image, a complex PolSAR scene is partitioned into aggregated, structural and homogenous pixel-level subspaces with the characteristics of relatively coherent terrain types in each subspace. Then, according to the characteristics of three subspaces above, three specific methods are adopted, and furthermore polarimetric information is exploited to improve the segmentation result. Experimental results on PolSAR data sets with different bands and sensors demonstrate that the proposed method is superior to the state-of-the-art methods in region homogeneity and edge preservation for terrain classification.
We propose a novel approach to identify one of the most significant dermoscopic criteria in the diagnosis of Cutaneous Melanoma: the Blue-whitish structure. In this paper, we achieve this goal in a Multiple Instance Learning framework using only image-level labels of whether the feature is present or not. As the output, we predict the image classification label and as well localize the feature in the image. Experiments are conducted on a challenging dataset with results outperforming state-of-the-art. This study provides an improvement on the scope of modelling for computerized image analysis of skin lesions, in particular in that it puts forward a framework for identification of dermoscopic local features from weakly-labelled data.
In this paper, we propose a novel method to find characteristic landmarks on ancient Roman imperial coins using deep convolutional neural network models (CNNs). We formulate an optimization problem to discover class-specific regions while guaranteeing specific controlled loss of accuracy. Analysis on visualization of the discovered region confirms that not only can the proposed method successfully find a set of characteristic regions per class, but also the discovered region is consistent with human expert annotations. We also propose a new framework to recognize the Roman coins which exploits hierarchical structure of the ancient Roman coins using the state-of-the-art classification power of the CNNs adopted to a new task of coin classification. Experimental results show that the proposed framework is able to effectively recognize the ancient Roman coins. For this research, we have collected a new Roman coin dataset where all coins are annotated and consist of observe (head) and reverse (tail) images.
Single Index Models (SIMs) are simple yet flexible semi-parametric models for classification and regression. Response variables are modeled as a nonlinear, monotonic function of a linear combination of features. Estimation in this context requires learning both the feature weights, and the nonlinear function. While methods have been described to learn SIMs in the low dimensional regime, a method that can efficiently learn SIMs in high dimensions has not been forthcoming. We propose three variants of a computationally and statistically efficient algorithm for SIM inference in high dimensions. We establish excess risk bounds for the proposed algorithms and experimentally validate the advantages that our SIM learning methods provide relative to Generalized Linear Model (GLM) and low dimensional SIM based learning methods.
This paper investigates the problem of active learning for binary label prediction on a graph. We introduce a simple and label-efficient algorithm called S2 for this task. At each step, S2 selects the vertex to be labeled based on the structure of the graph and all previously gathered labels. Specifically, S2 queries for the label of the vertex that bisects the *shortest shortest* path between any pair of oppositely labeled vertices. We present a theoretical estimate of the number of queries S2 needs in terms of a novel parametrization of the complexity of binary functions on graphs. We also present experimental results demonstrating the performance of S2 on both real and synthetic data. While other graph-based active learning algorithms have shown promise in practice, our algorithm is the first with both good performance and theoretical guarantees. Finally, we demonstrate the implications of the S2 algorithm to the theory of nonparametric active learning. In particular, we show that S2 achieves near minimax optimal excess risk for an important class of nonparametric classification problems.
Quantitative analysis of channel networks plays an important role in river studies. To provide a quantitative representation of channel networks, we propose a new method that extracts channels from remotely sensed images and estimates their widths. Our fully automated method is based on a recently proposed Multiscale Singularity Index that responds strongly to curvilinear structures but weakly to edges. The algorithm produces a channel map, using a single image where water and non-water pixels have contrast, such as a Landsat near-infrared band image or a water index defined on multiple bands. The proposed method provides a robust alternative to the procedures that are used in remote sensing of fluvial geomorphology and makes classification and analysis of channel networks easier. The source code of the algorithm is available at: http://live.ece.utexas.edu/research/cne/.