The World Health Organization (WHO) reported 1.25 million deaths yearly due to road traffic accidents worldwide and the number has been continuously increasing over the last few years. Nearly fifth of these accidents are caused by distracted drivers. Existing work of distracted driver detection is concerned with a small set of distractions (mostly, cell phone usage). Unreliable ad-hoc methods are often used.In this paper, we present the first publicly available dataset for driver distraction identification with more distraction postures than existing alternatives. In addition, we propose a reliable deep learning-based solution that achieves a 90% accuracy. The system consists of a genetically-weighted ensemble of convolutional neural networks, we show that a weighted ensemble of classifiers using a genetic algorithm yields in a better classification confidence. We also study the effect of different visual elements in distraction detection by means of face and hand localizations, and skin segmentation. Finally, we present a thinned version of our ensemble that could achieve 84.64% classification accuracy and operate in a real-time environment.
Normal Pressure Hydrocephalus (NPH) is one of the few reversible forms of dementia, Due to their low cost and versatility, Computed Tomography (CT) scans have long been used as an aid to help diagnose intracerebral anomalies such as NPH. However, no well-defined and effective protocol currently exists for the analysis of CT scan-based ventricular, cerebral mass and subarachnoid space volumes in the setting of NPH. The Evan's ratio, an approximation of the ratio of ventricle to brain volume using only one 2D slice of the scan, has been proposed but is not robust. Instead of manually measuring a 2-dimensional proxy for the ratio of ventricle volume to brain volume, this study proposes an automated method of calculating the brain volumes for better recognition of NPH from a radiological standpoint. The method first aligns the subject CT volume to a common space through an affine transformation, then uses a random forest classifier to mask relevant tissue types. A 3D morphological segmentation method is used to partition the brain volume, which in turn is used to train machine learning methods to classify the subjects into non-NPH vs. NPH based on volumetric information. The proposed algorithm has increased sensitivity compared to the Evan's ratio thresholding method.
Two things seem to be indisputable in the contemporary deep learning discourse: 1. The categorical cross-entropy loss after softmax activation is the method of choice for classification. 2. Training a CNN classifier from scratch on small datasets does not work well. In contrast to this, we show that the cosine loss function provides significantly better performance than cross-entropy on datasets with only a handful of samples per class. For example, the accuracy achieved on the CUB-200-2011 dataset without pre-training is by 30% higher than with the cross-entropy loss. Further experiments on other popular datasets confirm our findings. Moreover, we demonstrate that integrating prior knowledge in the form of class hierarchies is straightforward with the cosine loss and improves classification performance further.
Music genre recognition based on visual representation has been successfully explored over the last years. Recently, there has been increasing interest in attempting convolutional neural networks (CNNs) to achieve the task. However, most of existing methods employ the mature CNN structures proposed in image recognition without any modification, which results in the learning features that are not adequate for music genre classification. Faced with the challenge of this issue, we fully exploit the low-level information from spectrograms of audios and develop a novel CNN architecture in this paper. The proposed CNN architecture takes the long contextual information into considerations, which transfers more suitable information for the decision-making layer. Various experiments on several benchmark datasets, including GTZAN, Ballroom, and Extended Ballroom, have verified the excellent performances of the proposed neural network. Codes and model will be available at "ttps://github.com/CaifengLiu/music-genre-classification".
The aim of this paper was to compare soft confusion matrix approach and Bayes metaclassifier under the multi-label classification framework. Although the methods were successfully applied under the multi-label classification framework, they have not been compared directly thus far. Such comparison is of vital importance because both methods are quite similar as they are both based on the concept of randomized reference classifier. Since both algorithms were designed to deal with single-label problems, they are combined with the problem-transformation approach to multi-label classification. Present study included 29 benchmark datasets and four different base classifiers. The algorithms were compared in terms of 11 quality criteria and the results were subjected to statistical analysis.
Approximate vanishing ideal is a concept from computer algebra that studies the algebraic varieties behind perturbed data points. To capture the nonlinear structure of perturbed points, the introduction of approximation to exact vanishing ideals plays a critical role. However, such an approximation also gives rise to a theoretical problem---the spurious vanishing problem---in the basis construction of approximate vanishing ideals; namely, obtained basis polynomials can be approximately vanishing simply because of the small coefficients. In this paper, we propose a first general method that enables various basis construction algorithms to overcome the spurious vanishing problem. In particular, we integrate coefficient normalization with polynomial-based basis constructions, which do not need the proper ordering of monomials to process for basis constructions. We further propose a method that takes advantage of the iterative nature of basis construction so that computationally costly operations for coefficient normalization can be circumvented. Moreover, a coefficient truncation method is proposed for further accelerations. From the experiments, it can be shown that the proposed method overcomes the spurious vanishing problem, resulting in shorter feature vectors while sustaining comparable or even lower classification error.
YouTube is the leading social media platform for sharing videos. As a result, it is plagued with misleading content that includes staged videos presented as real footages from an incident, videos with misrepresented context and videos where audio/video content is morphed. We tackle the problem of detecting such misleading videos as a supervised classification task. We develop UCNet - a deep network to detect fake videos and perform our experiments on two datasets - VAVD created by us and publicly available FVC [8]. We achieve a macro averaged F-score of 0.82 while training and testing on a 70:30 split of FVC, while the baseline model scores 0.36. We find that the proposed model generalizes well when trained on one dataset and tested on the other.
The remarkable success of machine learning, especially deep learning, has produced a variety of cloud-based services for mobile users. Such services require an end user to send data to the service provider, which presents a serious challenge to end-user privacy. To address this concern, prior works either add noise to the data or send features extracted from the raw data. They struggle to balance between the utility and privacy because added noise reduces utility and raw data can be reconstructed from extracted features. This work represents a methodical departure from prior works: we balance between a measure of privacy and another of utility by leveraging adversarial learning to find a sweeter tradeoff. We design an encoder that optimizes against the reconstruction error (a measure of privacy), adversarially by a Decoder, and the inference accuracy (a measure of utility) by a Classifier. The result is RAN, a novel deep model with a new training algorithm that automatically extracts features for classification that are both private and useful. It turns out that adversarially forcing the extracted features to only conveys the intended information required by classification leads to an implicit regularization leading to better classification accuracy than the original model which completely ignores privacy. Thus, we achieve better privacy with better utility, a surprising possibility in machine learning! We conducted extensive experiments on five popular datasets over four training schemes, and demonstrate the superiority of RAN compared with existing alternatives.
Many real-world problems exhibit the coexistence of multiple types of heterogeneity, such as view heterogeneity (i.e., multi-view property) and task heterogeneity (i.e., multi-task property). For example, in an image classification problem containing multiple poses of the same object, each pose can be considered as one view, and the detection of each type of object can be treated as one task. Furthermore, in some problems, the data type of multiple views might be different. In a web classification problem, for instance, we might be provided an image and text mixed data set, where the web pages are characterized by both images and texts. A common strategy to solve this kind of problem is to leverage the consistency of views and the relatedness of tasks to build the prediction model. In the context of deep neural network, multi-task relatedness is usually realized by grouping tasks at each layer, while multi-view consistency is usually enforced by finding the maximal correlation coefficient between views. However, there is no existing deep learning algorithm that jointly models task and view dual heterogeneity, particularly for a data set with multiple modalities (text and image mixed data set or text and video mixed data set, etc.). In this paper, we bridge this gap by proposing a deep multi-task multi-view learning framework that learns a deep representation for such dual-heterogeneity problems. Empirical studies on multiple real-world data sets demonstrate the effectiveness of our proposed Deep-MTMV algorithm.
We present a novel Convolutional Neural Network (CNN) based approach for one class classification. The idea is to use a zero centered Gaussian noise in the latent space as the pseudo-negative class and train the network using the cross-entropy loss to learn a good representation as well as the decision boundary for the given class. A key feature of the proposed approach is that any pre-trained CNN can be used as the base network for one class classification. The proposed One Class CNN (OC-CNN) is evaluated on the UMDAA-02 Face, Abnormality-1001, FounderType-200 datasets. These datasets are related to a variety of one class application problems such as user authentication, abnormality detection and novelty detection. Extensive experiments demonstrate that the proposed method achieves significant improvements over the recent state-of-the-art methods. The source code is available at : github.com/otkupjnoz/oc-cnn.
We address two questions for training a convolutional neural network (CNN) for hyperspectral image classification: i) is it possible to build a pre-trained network? and ii) is the pre-training effective in furthering the performance? To answer the first question, we have devised an approach that pre-trains a network on multiple source datasets that differ in their hyperspectral characteristics and fine-tunes on a target dataset. This approach effectively resolves the architectural issue that arises when transferring meaningful information between the source and the target networks. To answer the second question, we carried out several ablation experiments. Based on the experimental results, a network trained from scratch performs as good as a network fine-tuned from a pre-trained network. However, we observed that pre-training the network has its own advantage in achieving better performances when deeper networks are required.
Evolutionary computation (EC) algorithms, such as discrete and multi-objective versions of particle swarm optimization (PSO), have been applied to solve the Feature selection (FS) problem, tackling the combinatorial explosion of search spaces that are peppered with local minima. Furthermore, high-dimensional FS problems such as finding a small set of biomarkers to make a diagnostic call add an additional challenge as such methods ability to pick out the most important features must remain unchanged in decision spaces of increasing dimensions and presence of irrelevant features. We developed a combinatorial PSO algorithm, called COMB-PSO, that scales up to high-dimensional gene expression data while still selecting the smallest subsets of genes that allow reliable classification of samples. In particular, COMB-PSO enhances the encoding, speed of convergence, control of divergence and diversity of the conventional PSO algorithm, balancing exploration and exploitation of the search space. Applying our approach on real gene expression data of different cancers, COMB-PSO finds gene sets of smallest size that allow a reliable classification of the underlying disease classes.
We employ triplet loss as a feature embedding regularizer to boost classification performance. Standard architectures, like ResNet and Inception, are extended to support both losses with minimal hyper-parameter tuning. This promotes generality while fine-tuning pretrained networks. Triplet loss is a powerful surrogate for recently proposed embedding regularizers. Yet, it is avoided due to large batch-size requirement and high computational cost. Through our experiments, we re-assess these assumptions. During inference, our network supports both classification and embedding tasks without any computational overhead. Quantitative evaluation highlights a steady improvement on five fine-grained recognition datasets. Further evaluation on an imbalanced video dataset achieves significant improvement. Triplet loss brings feature embedding characteristics like nearest neighbor to classification models. Code available at \url{http://bit.ly/2LNYEqL}.
We study the neural-linear bandit model for solving sequential decision-making problems with high dimensional side information. Neural-linear bandits leverage the representation power of deep neural networks and combine it with efficient exploration mechanisms, designed for linear contextual bandits, on top of the last hidden layer. Since the representation is being optimized during learning, information regarding exploration with "old" features is lost. Here, we propose the first limited memory neural-linear bandit that is resilient to this phenomenon, which we term catastrophic forgetting. We evaluate our method on a variety of real-world data sets, including regression, classification, and sentiment analysis, and observe that our algorithm is resilient to catastrophic forgetting and achieves superior performance.
Colorectal liver metastasis is one of most aggressive liver malignancies. While the definition of lesion type based on CT images determines the diagnosis and therapeutic strategy, the discrimination between cancerous and non-cancerous lesions are critical and requires highly skilled expertise, experience and time. In the present work we introduce an end-to-end deep learning approach to assist in the discrimination between liver metastases from colorectal cancer and benign cysts in abdominal CT images of the liver. Our approach incorporates the efficient feature extraction of InceptionV3 combined with residual connections and pre-trained weights from ImageNet. The architecture also includes fully connected classification layers to generate a probabilistic output of lesion type. We use an in-house clinical biobank with 230 liver lesions originating from 63 patients. With an accuracy of 0.96 and a F1-score of 0.92, the results obtained with the proposed approach surpasses state of the art methods. Our work provides the basis for incorporating machine learning tools in specialized radiology software to assist physicians in the early detection and treatment of liver lesions.
We present a large-scale study of gender bias in occupation classification, a task where the use of machine learning may lead to negative outcomes on peoples' lives. We analyze the potential allocation harms that can result from semantic representation bias. To do so, we study the impact on occupation classification of including explicit gender indicators---such as first names and pronouns---in different semantic representations of online biographies. Additionally, we quantify the bias that remains when these indicators are "scrubbed," and describe proxy behavior that occurs in the absence of explicit gender indicators. As we demonstrate, differences in true positive rates between genders are correlated with existing gender imbalances in occupations, which may compound these imbalances.
We present ADMM-Softmax, an alternating direction method of multipliers (ADMM) for solving multinomial logistic regression (MLR) problems. Our method is geared toward supervised classification tasks with many examples and features. It decouples the nonlinear optimization problem in MLR into three steps that can be solved efficiently. In particular, each iteration of ADMM-Softmax consists of a linear least-squares problem, a set of independent small-scale smooth, convex problems, and a trivial dual variable update. Solution of the least-squares problem can be be accelerated by pre-computing a factorization or preconditioner, and the separability in the smooth, convex problem can be easily parallelized across examples. For two image classification problems, we demonstrate that ADMM-Softmax leads to improved generalization compared to a Newton-Krylov, a quasi Newton, and a stochastic gradient descent method.
Background: The trend towards large-scale studies including population imaging poses new challenges in terms of quality control (QC). This is a particular issue when automatic processing tools, e.g. image segmentation methods, are employed to derive quantitative measures or biomarkers for later analyses. Manual inspection and visual QC of each segmentation isn't feasible at large scale. However, it's important to be able to automatically detect when a segmentation method fails so as to avoid inclusion of wrong measurements into subsequent analyses which could lead to incorrect conclusions. Methods: To overcome this challenge, we explore an approach for predicting segmentation quality based on Reverse Classification Accuracy, which enables us to discriminate between successful and failed segmentations on a per-cases basis. We validate this approach on a new, large-scale manually-annotated set of 4,800 cardiac magnetic resonance scans. We then apply our method to a large cohort of 7,250 cardiac MRI on which we have performed manual QC. Results: We report results used for predicting segmentation quality metrics including Dice Similarity Coefficient (DSC) and surface-distance measures. As initial validation, we present data for 400 scans demonstrating 99% accuracy for classifying low and high quality segmentations using predicted DSC scores. As further validation we show high correlation between real and predicted scores and 95% classification accuracy on 4,800 scans for which manual segmentations were available. We mimic real-world application of the method on 7,250 cardiac MRI where we show good agreement between predicted quality metrics and manual visual QC scores. Conclusions: We show that RCA has the potential for accurate and fully automatic segmentation QC on a per-case basis in the context of large-scale population imaging as in the UK Biobank Imaging Study.
We introduce the concrete autoencoder, an end-to-end differentiable method for global feature selection, which efficiently identifies a subset of the most informative features and simultaneously learns a neural network to reconstruct the input data from the selected features. Our method is unsupervised, and is based on using a concrete selector layer as the encoder and using a standard neural network as the decoder. During the training phase, the temperature of the concrete selector layer is gradually decreased, which encourages a user-specified number of discrete features to be learned. During test time, the selected features can be used with the decoder network to reconstruct the remaining input features. We evaluate concrete autoencoders on a variety of datasets, where they significantly outperform state-of-the-art methods for feature selection and data reconstruction. In particular, on a large-scale gene expression dataset, the concrete autoencoder selects a small subset of genes whose expression levels can be use to impute the expression levels of the remaining genes. In doing so, it improves on the current widely-used expert-curated L1000 landmark genes, potentially reducing measurement costs by 20%. The concrete autoencoder can be implemented by adding just a few lines of code to a standard autoencoder.
Normalization layers are a staple in state-of-the-art deep neural network architectures. They are widely believed to stabilize training, enable higher learning rate, accelerate convergence and improve generalization, though the reason for their effectiveness is still an active research topic. In this work, we challenge the commonly-held beliefs by showing that none of the perceived benefits is unique to normalization. Specifically, we propose fixed-update initialization (Fixup), an initialization motivated by solving the exploding and vanishing gradient problem at the beginning of training via properly rescaling a standard initialization. We find training residual networks with Fixup to be as stable as training with normalization -- even for networks with 10,000 layers. Furthermore, with proper regularization, Fixup enables residual networks without normalization to achieve state-of-the-art performance in image classification and machine translation.