Purpose: Deep neural networks (DNNs) have been widely applied in medical image classification, benefiting from its powerful mapping capability among medical images. However, these existing deep learning-based methods depend on an enormous amount of carefully labeled images. Meanwhile, noise is inevitably introduced in the labeling process, degrading the performance of models. Hence, it's significant to devise robust training strategies to mitigate label noise in the medical image classification tasks. Methods: In this work, we propose a novel Bayesian statistics guided label refurbishment mechanism (BLRM) for DNNs to prevent overfitting noisy images. BLRM utilizes maximum a posteriori probability (MAP) in the Bayesian statistics and the exponentially time-weighted technique to selectively correct the labels of noisy images. The training images are purified gradually with the training epochs when BLRM is activated, further improving classification performance. Results: Comprehensive experiments on both synthetic noisy images (public OCT & Messidor datasets) and real-world noisy images (ANIMAL-10N) demonstrate that BLRM refurbishes the noisy labels selectively, curbing the adverse effects of noisy data. Also, the anti-noise BLRM integrated with DNNs are effective at different noise ratio and are independent of backbone DNN architectures. In addition, BLRM is superior to state-of-the-art comparative methods of anti-noise. Conclusions: These investigations indicate that the proposed BLRM is well capable of mitigating label noise in medical image classification tasks.
We introduce Transductive Infomation Maximization (TIM) for few-shot learning. Our method maximizes the mutual information between the query features and their label predictions for a given few-shot task, in conjunction with a supervision loss based on the support set. We motivate our transductive loss by deriving a formal relation between the classification accuracy and mutual-information maximization. Furthermore, we propose a new alternating-direction solver, which substantially speeds up transductive inference over gradient-based optimization, while yielding competitive accuracy. We also provide a convergence analysis of our solver based on Zangwill's theory and bound-optimization arguments. TIM inference is modular: it can be used on top of any base-training feature extractor. Following standard transductive few-shot settings, our comprehensive experiments demonstrate that TIM outperforms state-of-the-art methods significantly across various datasets and networks, while used on top of a fixed feature extractor trained with simple cross-entropy on the base classes, without resorting to complex meta-learning schemes. It consistently brings between 2 % and 5 % improvement in accuracy over the best performing method, not only on all the well-established few-shot benchmarks but also on more challenging scenarios, with random tasks, domain shift and larger numbers of classes, as in the recently introduced META-DATASET. Our code is publicly available at https://github.com/mboudiaf/TIM. We also publicly release a standalone PyTorch implementation of META-DATASET, along with additional benchmarking results, at https://github.com/mboudiaf/pytorch-meta-dataset.
The natural world is long-tailed: rare classes are observed orders of magnitudes less frequently than common ones, leading to highly-imbalanced data where rare classes can have only handfuls of examples. Learning from few examples is a known challenge for deep learning based classification algorithms, and is the focus of the field of low-shot learning. One potential approach to increase the training data for these rare classes is to augment the limited real data with synthetic samples. This has been shown to help, but the domain shift between real and synthetic hinders the approaches' efficacy when tested on real data. We explore the use of image-to-image translation methods to close the domain gap between synthetic and real imagery for animal species classification in data collected from camera traps: motion-activated static cameras used to monitor wildlife. We use low-level feature alignment between source and target domains to make synthetic data for a rare species generated using a graphics engine more "realistic". Compared against a system augmented with unaligned synthetic data, our experiments show a considerable decrease in classification error rates on a rare species.
As predictive models are deployed into the real world, they must increasingly contend with strategic behavior. A growing body of work on strategic classification treats this problem as a Stackelberg game: the decision-maker "leads" in the game by deploying a model, and the strategic agents "follow" by playing their best response to the deployed model. Importantly, in this framing, the burden of learning is placed solely on the decision-maker, while the agents' best responses are implicitly treated as instantaneous. In this work, we argue that the order of play in strategic classification is fundamentally determined by the relative frequencies at which the decision-maker and the agents adapt to each other's actions. In particular, by generalizing the standard model to allow both players to learn over time, we show that a decision-maker that makes updates faster than the agents can reverse the order of play, meaning that the agents lead and the decision-maker follows. We observe in standard learning settings that such a role reversal can be desirable for both the decision-maker and the strategic agents. Finally, we show that a decision-maker with the freedom to choose their update frequency can induce learning dynamics that converge to Stackelberg equilibria with either order of play.
Intelligent systems are transforming the world, as well as our healthcare system. We propose a deep learning-based cough sound classification model that can distinguish between children with healthy versus pathological coughs such as asthma, upper respiratory tract infection (URTI), and lower respiratory tract infection (LRTI). In order to train a deep neural network model, we collected a new dataset of cough sounds, labelled with clinician's diagnosis. The chosen model is a bidirectional long-short term memory network (BiLSTM) based on Mel Frequency Cepstral Coefficients (MFCCs) features. The resulting trained model when trained for classifying two classes of coughs -- healthy or pathology (in general or belonging to a specific respiratory pathology), reaches accuracy exceeding 84\% when classifying cough to the label provided by the physicians' diagnosis. In order to classify subject's respiratory pathology condition, results of multiple cough epochs per subject were combined. The resulting prediction accuracy exceeds 91\% for all three respiratory pathologies. However, when the model is trained to classify and discriminate among the four classes of coughs, overall accuracy dropped: one class of pathological coughs are often misclassified as other. However, if one consider the healthy cough classified as healthy and pathological cough classified to have some kind of pathologies, then the overall accuracy of four class model is above 84\%. A longitudinal study of MFCC feature space when comparing pathological and recovered coughs collected from the same subjects revealed the fact that pathological cough irrespective of the underlying conditions occupy the same feature space making it harder to differentiate only using MFCC features.
Numerous malware families rely on domain generation algorithms (DGAs) to establish a connection to their command and control (C2) server. Counteracting DGAs, several machine learning classifiers have been proposed enabling the identification of the DGA that generated a specific domain name and thus triggering targeted remediation measures. However, the proposed state-of-the-art classifiers are based on deep learning models. The black box nature of these makes it difficult to evaluate their reasoning. The resulting lack of confidence makes the utilization of such models impracticable. In this paper, we propose EXPLAIN, a feature-based and contextless DGA multiclass classifier. We comparatively evaluate several combinations of feature sets and hyperparameters for our approach against several state-of-the-art classifiers in a unified setting on the same real-world data. Our classifier achieves competitive results, is real-time capable, and its predictions are easier to trace back to features than the predictions made by the DGA multiclass classifiers proposed in related work.
In multi-label classification, where a single example may be associated with several class labels at the same time, the ability to model dependencies between labels is considered crucial to effectively optimize non-decomposable evaluation measures, such as the Subset 0/1 loss. The gradient boosting framework provides a well-studied foundation for learning models that are specifically tailored to such a loss function and recent research attests the ability to achieve high predictive accuracy in the multi-label setting. The utilization of second-order derivatives, as used by many recent boosting approaches, helps to guide the minimization of non-decomposable losses, due to the information about pairs of labels it incorporates into the optimization process. On the downside, this comes with high computational costs, even if the number of labels is small. In this work, we address the computational bottleneck of such approach -- the need to solve a system of linear equations -- by integrating a novel approximation technique into the boosting procedure. Based on the derivatives computed during training, we dynamically group the labels into a predefined number of bins to impose an upper bound on the dimensionality of the linear system. Our experiments, using an existing rule-based algorithm, suggest that this may boost the speed of training, without any significant loss in predictive performance.
This paper is concerned with algorithms for prediction of discrete sequences over a finite alphabet, using variable order Markov models. The class of such algorithms is large and in principle includes any lossless compression algorithm. We focus on six prominent prediction algorithms, including Context Tree Weighting (CTW), Prediction by Partial Match (PPM) and Probabilistic Suffix Trees (PSTs). We discuss the properties of these algorithms and compare their performance using real life sequences from three domains: proteins, English text and music pieces. The comparison is made with respect to prediction quality as measured by the average log-loss. We also compare classification algorithms based on these predictors with respect to a number of large protein classification tasks. Our results indicate that a "decomposed" CTW (a variant of the CTW algorithm) and PPM outperform all other algorithms in sequence prediction tasks. Somewhat surprisingly, a different algorithm, which is a modification of the Lempel-Ziv compression algorithm, significantly outperforms all algorithms on the protein classification problems.
Recent advances in protecting node privacy on graph data and attacking graph neural networks (GNNs) gain much attention. The eye does not bring these two essential tasks together yet. Imagine an adversary can utilize the powerful GNNs to infer users' private labels in a social network. How can we adversarially defend against such privacy attacks while maintaining the utility of perturbed graphs? In this work, we propose a novel research task, adversarial defenses against GNN-based privacy attacks, and present a graph perturbation-based approach, NetFense, to achieve the goal. NetFense can simultaneously keep graph data unnoticeability (i.e., having limited changes on the graph structure), maintain the prediction confidence of targeted label classification (i.e., preserving data utility), and reduce the prediction confidence of private label classification (i.e., protecting the privacy of nodes). Experiments conducted on single- and multiple-target perturbations using three real graph data exhibit that the perturbed graphs by NetFense can effectively maintain data utility (i.e., model unnoticeability) on targeted label classification and significantly decrease the prediction confidence of private label classification (i.e., privacy protection). Extensive studies also bring several insights, such as the flexibility of NetFense, preserving local neighborhoods in data unnoticeability, and better privacy protection for high-degree nodes.
In many computer vision classification tasks, class priors at test time often differ from priors on the training set. In the case of such prior shift, classifiers must be adapted correspondingly to maintain close to optimal performance. This paper analyzes methods for adaptation of probabilistic classifiers to new priors and for estimating new priors on an unlabeled test set. We propose a novel method to address a known issue of prior estimation methods based on confusion matrices, where inconsistent estimates of decision probabilities and confusion matrices lead to negative values in the estimated priors. Experiments on fine-grained image classification datasets provide insight into the best practice of prior shift estimation and classifier adaptation, and show that the proposed method achieves state-of-the-art results in prior adaptation. Applying the best practice to two tasks with naturally imbalanced priors, learning from web-crawled images and plant species classification, increased the recognition accuracy by 1.1% and 3.4% respectively.
Image reconstruction is likely the most predominant auxiliary task for image classification, but we would like to think twice about this convention. In this paper, we investigated "approximating the Fourier Transform of the input image" as a potential alternative, in the hope that it may further boost the performances on the primary task or introduce novel constraints not well covered by image reconstruction. We experimented with five popular classification architectures on the CIFAR-10 dataset, and the empirical results indicated that our proposed auxiliary task generally improves the classification accuracy. More notably, the results showed that in certain cases our proposed auxiliary task may enhance the classifiers' resistance to adversarial attacks generated using the fast gradient sign method.
Most modern neural networks for classification fail to take into account the concept of the unknown. Trained neural networks are usually tested in an unrealistic scenario with only examples from a closed set of known classes. In an attempt to develop a more realistic model, the concept of working in an open set environment has been introduced. This in turn leads to the concept of incremental learning where a model with its own architecture and initial trained set of data can identify unknown classes during the testing phase and autonomously update itself if evidence of a new class is detected. Some problems that arise in incremental learning are inefficient use of resources to retrain the classifier repeatedly and the decrease of classification accuracy as multiple classes are added over time. This process of instantiating new classes is repeated as many times as necessary, accruing errors. To address these problems, this paper proposes the Classification Confidence Threshold approach to prime neural networks for incremental learning to keep accuracies high by limiting forgetting. A lean method is also used to reduce resources used in the retraining of the neural network. The proposed method is based on the idea that a network is able to incrementally learn a new class even when exposed to a limited number samples associated with the new class. This method can be applied to most existing neural networks with minimal changes to network architecture.
Multi-label image classification (MLIC) is a fundamental and practical task, which aims to assign multiple possible labels to an image. In recent years, many deep convolutional neural network (CNN) based approaches have been proposed which model label correlations to discover semantics of labels and learn semantic representations of images. This paper advances this research direction by improving both the modeling of label correlations and the learning of semantic representations. On the one hand, besides the local semantics of each label, we propose to further explore global semantics shared by multiple labels. On the other hand, existing approaches mainly learn the semantic representations at the last convolutional layer of a CNN. But it has been noted that the image representations of different layers of CNN capture different levels or scales of features and have different discriminative abilities. We thus propose to learn semantic representations at multiple convolutional layers. To this end, this paper designs a Multi-layered Semantic Representation Network (MSRN) which discovers both local and global semantics of labels through modeling label correlations and utilizes the label semantics to guide the semantic representations learning at multiple layers through an attention mechanism. Extensive experiments on four benchmark datasets including VOC 2007, COCO, NUS-WIDE, and Apparel show a competitive performance of the proposed MSRN against state-of-the-art models.
In this study, a natural language processing-based (NLP-based) method is proposed for the sector-wise automatic classification of ad texts created on online advertising platforms. Our data set consists of approximately 21,000 labeled advertising texts from 12 different sectors. In the study, the Bidirectional Encoder Representations from Transformers (BERT) model, which is a transformer-based language model that is recently used in fields such as text classification in the natural language processing literature, was used. The classification efficiencies obtained using a pre-trained BERT model for the Turkish language are shown in detail.
Compositional generalization is the ability to generalize systematically to a new data distribution by combining known components. Although humans seem to have a great ability to generalize compositionally, state-of-the-art neural models struggle to do so. In this work, we study compositional generalization in classification tasks and present two main contributions. First, we study ways to convert a natural language sequence-to-sequence dataset to a classification dataset that also requires compositional generalization. Second, we show that providing structural hints (specifically, providing parse trees and entity links as attention masks for a Transformer model) helps compositional generalization.
Visual Analytics (VA) tools and techniques have been instrumental in supporting users to build better classification models, interpret models' overall logic, and audit results. In a different direction, VA has recently been applied to transform classification models into descriptive mechanisms instead of predictive. The idea is to use such models as surrogates for data patterns, visualizing the model to understand the phenomenon represented by the data. Although very useful and inspiring, the few proposed approaches have opted to use low complex classification models to promote straightforward interpretation, presenting limitations to capture intricate data patterns. In this paper, we present VAX (multiVariate dAta eXplanation), a new VA method to support the identification and visual interpretation of patterns in multivariate datasets. Unlike the existing similar approaches, VAX uses the concept of Jumping Emerging Patterns to identify and aggregate several diversified patterns, producing explanations through logic combinations of data variables. The potential of VAX to interpret complex multivariate datasets is demonstrated through use cases employing two real-world datasets covering different scenarios.
The literature on "benign overfitting" in overparameterized models has been mostly restricted to regression or binary classification; however, modern machine learning operates in the multiclass setting. Motivated by this discrepancy, we study benign overfitting in multiclass linear classification. Specifically, we consider the following training algorithms on separable data: (i) empirical risk minimization (ERM) with cross-entropy loss, which converges to the multiclass support vector machine (SVM) solution; (ii) ERM with least-squares loss, which converges to the min-norm interpolating (MNI) solution; and, (iii) the one-vs-all SVM classifier. First, we provide a simple sufficient deterministic condition under which all three algorithms lead to classifiers that interpolate the training data and have equal accuracy. When the data is generated from Gaussian mixtures or a multinomial logistic model, this condition holds under high enough effective overparameterization. We also show that this sufficient condition is satisfied under "neural collapse", a phenomenon that is observed in training deep neural networks. Second, we derive novel bounds on the accuracy of the MNI classifier, thereby showing that all three training algorithms lead to benign overfitting under sufficient overparameterization. Ultimately, our analysis shows that good generalization is possible for SVM solutions beyond the realm in which typical margin-based bounds apply.
The grade of clear cell renal cell carcinoma (ccRCC) is a critical prognostic factor, making ccRCC nuclei grading a crucial task in RCC pathology analysis. Computer-aided nuclei grading aims to improve pathologists' work efficiency while reducing their misdiagnosis rate by automatically identifying the grades of tumor nuclei within histopathological images. Such a task requires precisely segment and accurately classify the nuclei. However, most of the existing nuclei segmentation and classification methods can not handle the inter-class similarity property of nuclei grading, thus can not be directly applied to the ccRCC grading task. In this paper, we propose a Composite High-Resolution Network for ccRCC nuclei grading. Specifically, we propose a segmentation network called W-Net that can separate the clustered nuclei. Then, we recast the fine-grained classification of nuclei to two cross-category classification tasks, based on two high-resolution feature extractors (HRFEs) which are proposed for learning these two tasks. The two HRFEs share the same backbone encoder with W-Net by a composite connection so that meaningful features for the segmentation task can be inherited for the classification task. Last, a head-fusion block is applied to generate the predicted label of each nucleus. Furthermore, we introduce a dataset for ccRCC nuclei grading, containing 1000 image patches with 70945 annotated nuclei. We demonstrate that our proposed method achieves state-of-the-art performance compared to existing methods on this large ccRCC grading dataset.
Existing computer vision research in categorization struggles with fine-grained attributes recognition due to the inherently high intra-class variances and low inter-class variances. SOTA methods tackle this challenge by locating the most informative image regions and rely on them to classify the complete image. The most recent work, Vision Transformer (ViT), shows its strong performance in both traditional and fine-grained classification tasks. In this work, we propose a multi-stage ViT framework for fine-grained image classification tasks, which localizes the informative image regions without requiring architectural changes using the inherent multi-head self-attention mechanism. We also introduce attention-guided augmentations for improving the model's capabilities. We demonstrate the value of our approach by experimenting with four popular fine-grained benchmarks: CUB-200-2011, Stanford Cars, Stanford Dogs, and FGVC7 Plant Pathology. We also prove our model's interpretability via qualitative results.
We study how to evaluate the quantitative information content of a region within an image for a particular label. To this end, we bridge class activation maps with information theory. We develop an informative class activation map (infoCAM). Given a classification task, infoCAM depict how to accumulate information of partial regions to that of the entire image toward a label. Thus, we can utilise infoCAM to locate the most informative features for a label. When applied to an image classification task, infoCAM performs better than the traditional classification map in the weakly supervised object localisation task. We achieve state-of-the-art results on Tiny-ImageNet.