This paper presents a novel fusion of low-level approaches for dimensionality reduction into an effective approach for high-level objects in neuromorphic camera data called Inceptive Event Time-Surfaces (IETS). IETSs overcome several limitations of conventional time-surfaces by increasing robustness to noise, promoting spatial consistency, and improving the temporal localization of (moving) edges. Combining IETS with transfer learning improves state-of-the-art performance on the challenging problem of object classification utilizing event camera data.
This paper describes one objective function for learning semantically coherent feature embeddings in multi-output classification problems, i.e., when the response variables have dimension higher than one. In particular, we consider the problems of identity retrieval and soft biometrics labelling in visual surveillance environments, which have been attracting growing interests. Inspired by the triplet loss [34] function, we propose a generalization that: 1) defines a metric that considers the number of agreeing labels between pairs of elements; and 2) disregards the notion of anchor, replacing d(A1, A2) < d(A1, B) by d(A, B) < d(C, D), for A, B, C, D distance constraints, according to the number of agreeing labels between pairs. As the triplet loss formulation, our proposal also privileges small distances between positive pairs, but at the same time explicitly enforces that the distance between other pairs corresponds directly to their similarity in terms of agreeing labels. This yields feature embeddings with a strong correspondence between the classes centroids and their semantic descriptions, i.e., where elements are closer to others that share some of their labels than to elements with fully disjoint labels membership. As practical effect, the proposed loss can be seen as particularly suitable for performing joint coarse (soft label) + fine (ID) inference, based on simple rules as k-neighbours, which is a novelty with respect to previous related loss functions. Also, in opposition to its triplet counterpart, the proposed loss is agnostic with regard to any demanding criteria for mining learning instances (such as the semi-hard pairs). Our experiments were carried out in five different datasets (BIODI, LFW, IJB-A, Megaface and PETA) and validate our assumptions, showing highly promising results.
The assignment flow recently introduced in the J. Math. Imaging and Vision 58/2 (2017), constitutes a high-dimensional dynamical system that evolves on an elementary statistical manifold and performs contextual labeling (classification) of data given in any metric space. Vertices of a given graph index the data points and define a system of neighborhoods. These neighborhoods together with nonnegative weight parameters define regularization of the evolution of label assignments to data points, through geometric averaging induced by the affine e-connection of information geometry. Regarding evolutionary game dynamics, the assignment flow may be characterized as a large system of replicator equations that are coupled by geometric averaging. This paper establishes conditions on the weight parameters that guarantee convergence of the continuous-time assignment flow to integral assignments (labelings), up to a negligible subset of situations that will not be encountered when working with real data in practice. Furthermore, we classify attractors of the flow and quantify corresponding basins of attraction. This provides convergence guarantees for the assignment flow which are extended to the discrete-time assignment flow that results from applying a Runge-Kutta-Munthe-Kaas scheme for numerical geometric integration of the assignment flow. Several counter-examples illustrate that violating the conditions may entail unfavorable behavior of the assignment flow regarding contextual data classification.
The problem of training with a small set of positive samples is known as few-shot learning (FSL). It is widely known that traditional deep learning (DL) algorithms usually show very good performance when trained with large datasets. However, in many applications, it is not possible to obtain such a high number of samples. In the image domain, typical FSL applications include those related to face recognition. In the audio domain, music fraud or speaker recognition can be clearly benefited from FSL methods. This paper deals with the application of FSL to the detection of specific and intentional acoustic events given by different types of sound alarms, such as door bells or fire alarms, using a limited number of samples. These sounds typically occur in domestic environments where many events corresponding to a wide variety of sound classes take place. Therefore, the detection of such alarms in a practical scenario can be considered an open-set recognition (OSR) problem. To address the lack of a dedicated public dataset for audio FSL, researchers usually make modifications on other available datasets. This paper is aimed at poviding the audio recognition community with a carefully annotated dataset (https://zenodo.org/record/3689288) for FSL in an OSR context comprised of 1360 clips from 34 classes divided into pattern sounds} and unwanted sounds. To facilitate and promote research on this area, results with state-of-the-art baseline systems based on transfer learning are also presented.
We consider a high-dimensional mixture of two Gaussians in the noisy regime where even an oracle knowing the centers of the clusters misclassifies a small but finite fraction of the points. We provide a rigorous analysis of the generalization error of regularized convex classifiers, including ridge, hinge and logistic regression, in the high-dimensional limit where the number $n$ of samples and their dimension $d$ go to infinity while their ratio is fixed to $\alpha= n/d$. We discuss surprising effects of the regularization that in some cases allows to reach the Bayes-optimal performances. We also illustrate the interpolation peak at low regularization, and analyze the role of the respective sizes of the two clusters.
Many classification problems focus on maximizing the performance only on the samples with the highest relevance instead of all samples. As an example, we can mention ranking problems, accuracy at the top or search engines where only the top few queries matter. In our previous work, we derived a general framework including several classes of these linear classification problems. In this paper, we extend the framework to nonlinear classifiers. Utilizing a similarity to SVM, we dualize the problems, add kernels and propose a componentwise dual ascent method.
Text classification is one of the most important and fundamental tasks in natural language processing. Performance of this task mainly dependents on text representation learning. Currently, most existing learning frameworks mainly focus on encoding local contextual information between words. These methods always neglect to exploit global clues, such as label information, for encoding text information. In this study, we propose a label-guided learning framework LguidedLearn for text representation and classification. Our method is novel but simple that we only insert a label-guided encoding layer into the commonly used text representation learning schemas. That label-guided layer performs label-based attentive encoding to map the universal text embedding (encoded by a contextual information learner) into different label spaces, resulting in label-wise embeddings. In our proposed framework, the label-guided layer can be easily and directly applied with a contextual encoding method to perform jointly learning. Text information is encoded based on both the local contextual information and the global label clues. Therefore, the obtained text embeddings are more robust and discriminative for text classification. Extensive experiments are conducted on benchmark datasets to illustrate the effectiveness of our proposed method.
Imputation of missing data is a common application in various classification problems where the feature training matrix has missingness. A widely used solution to this imputation problem is based on the lazy learning technique, $k$-nearest neighbor (kNN) approach. However, most of the previous work on missing data does not take into account the presence of the class label in the classification problem. Also, existing kNN imputation methods use variants of Minkowski distance as a measure of distance, which does not work well with heterogeneous data. In this paper, we propose a novel iterative kNN imputation technique based on class weighted grey distance between the missing datum and all the training data. Grey distance works well in heterogeneous data with missing instances. The distance is weighted by Mutual Information (MI) which is a measure of feature relevance between the features and the class label. This ensures that the imputation of the training data is directed towards improving classification performance. This class weighted grey kNN imputation algorithm demonstrates improved performance when compared to other kNN imputation algorithms, as well as standard imputation algorithms such as MICE and missForest, in imputation and classification problems. These problems are based on simulated scenarios and UCI datasets with various rates of missingness.
Stochastic gradient descent (SGD) with constant momentum and its variants such as Adam are the optimization algorithms of choice for training deep neural networks (DNNs). Since DNN training is incredibly computationally expensive, there is great interest in speeding up the convergence. Nesterov accelerated gradient (NAG) improves the convergence rate of gradient descent (GD) for convex optimization using a specially designed momentum; however, it accumulates error when an inexact gradient is used (such as in SGD), slowing convergence at best and diverging at worst. In this paper, we propose Scheduled Restart SGD (SRSGD), a new NAG-style scheme for training DNNs. SRSGD replaces the constant momentum in SGD by the increasing momentum in NAG but stabilizes the iterations by resetting the momentum to zero according to a schedule. Using a variety of models and benchmarks for image classification, we demonstrate that, in training DNNs, SRSGD significantly improves convergence and generalization; for instance in training ResNet200 for ImageNet classification, SRSGD achieves an error rate of 20.93% vs. the benchmark of 22.13%. These improvements become more significant as the network grows deeper. Furthermore, on both CIFAR and ImageNet, SRSGD reaches similar or even better error rates with significantly fewer training epochs compared to the SGD baseline.
In many applications, data is collected in batches, some of which are corrupt or even adversarial. Recent work derived optimal robust algorithms for estimating discrete distributions in this setting. We consider a general framework of robust learning from batches, and determine the limits of both classification and distribution estimation over arbitrary, including continuous, domains. Building on these results, we derive the first robust agnostic computationally-efficient learning algorithms for piecewise-interval classification, and for piecewise-polynomial, monotone, log-concave, and gaussian-mixture distribution estimation.
Adversarial training has shown its ability in producing models that are robust to perturbations on the input data, but usually at the expense of decrease in the standard accuracy. To mitigate this issue, it is commonly believed that more training data will eventually help such adversarially robust models generalize better on the benign/unperturbed test data. In this paper, however, we challenge this conventional belief and show that more training data can hurt the generalization of adversarially robust models in the classification problems. We first investigate the Gaussian mixture classification with a linear loss and identify three regimes based on the strength of the adversary. In the weak adversary regime, more data improves the generalization of adversarially robust models. In the medium adversary regime, with more training data, the generalization loss exhibits a double descent curve, which implies the existence of an intermediate stage where more training data hurts the generalization. In the strong adversary regime, more data almost immediately causes the generalization error to increase. Then we move to the analysis of a two-dimensional classification problem with a 0-1 loss. We prove that more data always hurts the generalization performance of adversarially trained models with large perturbations. To complement our theoretical results, we conduct empirical studies on Gaussian mixture classification, support vector machines (SVMs), and linear regression.
Deep neural networks have demonstrated state-of-the-art performance on many classification tasks. However, they have no inherent capability to recognize when their predictions are wrong. There have been several efforts in the recent past to detect natural errors but the suggested mechanisms pose additional energy requirements. To address this issue, we propose an ensemble of classifiers at hidden layers to enable energy efficient detection of natural errors. In particular, we append Relevant-features based Auxiliary Cells (RACs) which are class specific binary linear classifiers trained on relevant features. The consensus of RACs is used to detect natural errors. Based on combined confidence of RACs, classification can be terminated early, thereby resulting in energy efficient detection. We demonstrate the effectiveness of our technique on various image classification datasets such as CIFAR-10, CIFAR-100 and Tiny-ImageNet.
Domain adaptation approaches seek to learn from a source domain and generalize it to an unseen target domain. At present, the state-of-the-art unsupervised domain adaptation approaches for subjective text classification problems leverage unlabeled target data along with labeled source data. In this paper, we propose a novel method for domain adaptation of single-task text classification problems based on a simple but effective idea of diversity-based generalization that does not require unlabeled target data but still matches the state-of-the-art in performance. Diversity plays the role of promoting the model to better generalize and be indiscriminate towards domain shift by forcing the model not to rely on same features for prediction. We apply this concept on the most explainable component of neural networks, the attention layer. To generate sufficient diversity, we create a multi-head attention model and infuse a diversity constraint between the attention heads such that each head will learn differently. We further expand upon our model by tri-training and designing a procedure with an additional diversity constraint between the attention heads of the tri-trained classifiers. Extensive evaluation using the standard benchmark dataset of Amazon reviews and a newly constructed dataset of Crisis events shows that our fully unsupervised method matches with the competing baselines that uses unlabeled target data. Our results demonstrate that machine learning architectures that ensure sufficient diversity can generalize better; encouraging future research to design ubiquitously usable learning models without using unlabeled target data.
Classification of whole-brain functional connectivity MRI data with convolutional neural networks (CNNs) has shown promise, but the complexity of these models impedes understanding of which aspects of brain activity contribute to classification. While visualization techniques have been developed to interpret CNNs, bias inherent in the method of encoding abstract input data, as well as the natural variance of deep learning models, detract from the accuracy of these techniques. We introduce a stochastic encoding method in an ensemble of CNNs to classify functional connectomes by gender. We applied our method to resting-state and task data from the UK BioBank, using two visualization techniques to measure the salience of three brain networks involved in task- and resting-states, and their interaction. To regress confounding factors such as head motion, age, and intracranial volume, we introduced a multivariate balancing algorithm to ensure equal distributions of such covariates between classes in our data. We achieved a final AUROC of 0.8459. We found that resting-state data classifies more accurately than task data, with the inner salience network playing the most important role of the three networks overall in classification of resting-state data and connections to the central executive network in task data.
Many binary classification problems minimize misclassification above (or below) a threshold. We show that instances of ranking problems, accuracy at the top or hypothesis testing may be written in this form. We propose a general framework to handle these classes of problems and show which known methods (both known and newly proposed) fall into this framework. We provide a theoretical analysis of this framework and mention selected possible pitfalls the methods may encounter. We suggest several numerical improvements including the implicit derivative and stochastic gradient descent. We provide an extensive numerical study. Based both on the theoretical properties and numerical experiments, we conclude the paper by suggesting which method should be used in which situation.
Lexical relations describe how concepts are semantically related, in the form of relation triples. The accurate prediction of lexical relations between concepts is challenging, due to the sparsity of patterns indicating the existence of such relations. We propose the Knowledge-Enriched Meta-Learning (KEML) framework to address the task of lexical relation classification. In KEML, the LKB-BERT (Lexical Knowledge Base-BERT) model is presented to learn concept representations from massive text corpora, with rich lexical knowledge injected by distant supervision. A probabilistic distribution of auxiliary tasks is defined to increase the model's ability to recognize different types of lexical relations. We further combine a meta-learning process over the auxiliary task distribution and supervised learning to train the neural lexical relation classifier. Experiments over multiple datasets show that KEML outperforms state-of-the-art methods.
We present PointAugment, a new auto-augmentation framework that automatically optimizes and augments point cloud samples to enrich the data diversity when we train a classification network. Different from existing auto-augmentation methods for 2D images, PointAugment is sample-aware and takes an adversarial learning strategy to jointly optimize an augmentor network and a classifier network, such that the augmentor can learn to produce augmented samples that best fit the classifier. Moreover, we formulate a learnable point augmentation function with a shape-wise transformation and a point-wise displacement, and carefully design loss functions to adopt the augmented samples based on the learning progress of the classifier. Extensive experiments also confirm PointAugment's effectiveness and robustness to improve the performance of various networks on shape classification and retrieval.
Insect monitoring is crucial for understanding the consequences of rapid ecological changes, but taxa identification currently requires tedious manual expert work and cannot be scaled-up efficiently. Deep convolutional neural networks (CNNs), provide a viable way to significantly increase the biomonitoring volumes. However, taxa abundances are typically very imbalanced and the amounts of training images for the rarest classes are simply too low for deep CNNs. As a result, the samples from the rare classes are often completely missed, while detecting them has biological importance. In this paper, we propose combining the trained deep CNN with one-class classifiers to improve the rare species identification. One-class classification models are traditionally trained with much fewer samples and they can provide a mechanism to indicate samples potentially belonging to the rare classes for human inspection. Our experiments confirm that the proposed approach may indeed support moving towards partial automation of the taxa identification task.
Nowadays, graph-structured data are increasingly used to model complex systems. Meanwhile, detecting anomalies from graph has become a vital research problem of pressing societal concerns. Anomaly detection is an unsupervised learning task of identifying rare data that differ from the majority. As one of the dominant anomaly detection algorithms, One Class Support Vector Machine has been widely used to detect outliers. However, those traditional anomaly detection methods lost their effectiveness in graph data. Since traditional anomaly detection methods are stable, robust and easy to use, it is vitally important to generalize them to graph data. In this work, we propose One Class Graph Neural Network (OCGNN), a one-class classification framework for graph anomaly detection. OCGNN is designed to combine the powerful representation ability of Graph Neural Networks along with the classical one-class objective. Compared with other baselines, OCGNN achieves significant improvements in extensive experiments.
The study of model bias and variance with respect to decision boundaries is critically important in supervised classification. There is generally a tradeoff between the two, as fine-tuning of the decision boundary of a classification model to accommodate more boundary training samples (i.e., higher model complexity) may improve training accuracy (i.e., lower bias) but hurt generalization against unseen data (i.e., higher variance). By focusing on just classification boundary fine-tuning and model complexity, it is difficult to reduce both bias and variance. To overcome this dilemma, we take a different perspective and investigate a new approach to handle inaccuracy and uncertainty in the training data labels, which are inevitable in many applications where labels are conceptual and labeling is performed by human annotators. The process of classification can be undermined by uncertainty in the labels of the training data; extending a boundary to accommodate an inaccurately labeled point will increase both bias and variance. Our novel method can reduce both bias and variance by estimating the pointwise label uncertainty of the training set and accordingly adjusting the training sample weights such that those samples with high uncertainty are weighted down and those with low uncertainty are weighted up. In this way, uncertain samples have a smaller contribution to the objective function of the model's learning algorithm and exert less pull on the decision boundary. In a real-world physical activity recognition case study, the data presents many labeling challenges, and we show that this new approach improves model performance and reduces model variance.