Increasing effort in brain image analysis has been dedicated to early diagnosis of Alzheimer's disease (AD) based on neuroimaging data. Most existing studies have been focusing on binary classification problems, e.g., distinguishing AD patients from normal control (NC) elderly or mild cognitive impairment (MCI) individuals from NC elderly. However, identifying individuals with AD and MCI, especially MCI individuals who will convert to AD (progressive MCI, pMCI), in a single setting, is needed to achieve the goal of early diagnosis of AD. In this paper, we propose a deep ordinal ranking model for distinguishing NC, stable MCI (sMCI), pMCI, and AD at an individual subject level, taking into account the inherent ordinal severity of brain degeneration caused by normal aging, MCI, and AD, rather than formulating the classification as a multi-category classification problem. The proposed deep ordinal ranking model focuses on the hippocampal morphology of individuals and learns informative and discriminative features automatically. Experiment results based on a large cohort of individuals from the Alzheimer's Disease Neuroimaging Initiative (ADNI) indicate that the proposed method can achieve better performance than traditional multi-category classification techniques using shape and radiomics features from structural magnetic resonance imaging (MRI) data.
We address the problem of multi-class classification in the case where the number of classes is very large. We propose a double sampling strategy on top of a multi-class to binary reduction strategy, which transforms the original multi-class problem into a binary classification problem over pairs of examples. The aim of the sampling strategy is to overcome the curse of long-tailed class distributions exhibited in majority of large-scale multi-class classification problems and to reduce the number of pairs of examples in the expanded data. We show that this strategy does not alter the consistency of the empirical risk minimization principle defined over the double sample reduction. Experiments are carried out on DMOZ and Wikipedia collections with 10,000 to 100,000 classes where we show the efficiency of the proposed approach in terms of training and prediction time, memory consumption, and predictive performance with respect to state-of-the-art approaches.
We present an approach to adaptively utilize deep neural networks in order to reduce the evaluation time on new examples without loss of accuracy. Rather than attempting to redesign or approximate existing networks, we propose two schemes that adaptively utilize networks. We first pose an adaptive network evaluation scheme, where we learn a system to adaptively choose the components of a deep network to be evaluated for each example. By allowing examples correctly classified using early layers of the system to exit, we avoid the computational time associated with full evaluation of the network. We extend this to learn a network selection system that adaptively selects the network to be evaluated for each example. We show that computational time can be dramatically reduced by exploiting the fact that many examples can be correctly classified using relatively efficient networks and that complex, computationally costly networks are only necessary for a small fraction of examples. We pose a global objective for learning an adaptive early exit or network selection policy and solve it by reducing the policy learning problem to a layer-by-layer weighted binary classification problem. Empirically, these approaches yield dramatic reductions in computational cost, with up to a 2.8x speedup on state-of-the-art networks from the ImageNet image recognition challenge with minimal (<1%) loss of top5 accuracy.
We show through case studies that it is easier to estimate the fundamental limits of data processing than to construct explicit algorithms to achieve those limits. Focusing on binary classification, data compression, and prediction under logarithmic loss, we show that in the finite space setting, when it is possible to construct an estimator of the limits with vanishing error with $n$ samples, it may require at least $n\ln n$ samples to construct an explicit algorithm to achieve the limits.
This paper aims at refined error analysis for binary classification using support vector machine (SVM) with Gaussian kernel and convex loss. Our first result shows that for some loss functions such as the truncated quadratic loss and quadratic loss, SVM with Gaussian kernel can reach the almost optimal learning rate, provided the regression function is smooth. Our second result shows that, for a large number of loss functions, under some Tsybakov noise assumption, if the regression function is infinitely smooth, then SVM with Gaussian kernel can achieve the learning rate of order $m^{-1}$, where $m$ is the number of samples.
Background: A universal unanswered question in neuroscience and machine learning is whether computers can decode the patterns of the human brain. Multi-Voxels Pattern Analysis (MVPA) is a critical tool for addressing this question. However, there are two challenges in the previous MVPA methods, which include decreasing sparsity and noise in the extracted features and increasing the performance of prediction. Methods: In overcoming mentioned challenges, this paper proposes Anatomical Pattern Analysis (APA) for decoding visual stimuli in the human brain. This framework develops a novel anatomical feature extraction method and a new imbalance AdaBoost algorithm for binary classification. Further, it utilizes an Error-Correcting Output Codes (ECOC) method for multiclass prediction. APA can automatically detect active regions for each category of the visual stimuli. Moreover, it enables us to combine homogeneous datasets for applying advanced classification. Results and Conclusions: Experimental studies on 4 visual categories (words, consonants, objects and scrambled photos) demonstrate that the proposed approach achieves superior performance to state-of-the-art methods.
Charcoal rot is a fungal disease that thrives in warm dry conditions and affects the yield of soybeans and other important agronomic crops worldwide. There is a need for robust, automatic and consistent early detection and quantification of disease symptoms which are important in breeding programs for the development of improved cultivars and in crop production for the implementation of disease control measures for yield protection. Current methods of plant disease phenotyping are predominantly visual and hence are slow and prone to human error and variation. There has been increasing interest in hyperspectral imaging applications for early detection of disease symptoms. However, the high dimensionality of hyperspectral data makes it very important to have an efficient analysis pipeline in place for the identification of disease so that effective crop management decisions can be made. The focus of this work is to determine the minimal number of most effective hyperspectral bands that can distinguish between healthy and diseased specimens early in the growing season. Healthy and diseased hyperspectral data cubes were captured at 3, 6, 9, 12, and 15 days after inoculation. We utilized inoculated and control specimens from 4 different genotypes. Each hyperspectral image was captured at 240 different wavelengths in the range of 383 to 1032 nm. We used a combination of genetic algorithm as an optimizer and support vector machines as a classifier for identification of maximally effective band combinations. A binary classification between healthy and infected samples using six selected band combinations obtained a classification accuracy of 97% and a F1 score of 0.97 for the infected class. The results demonstrated that these carefully chosen bands are more informative than RGB images, and could be used in a multispectral camera for remote identification of charcoal rot infection in soybean.
We conduct an empirical study of machine learning functionalities provided by major cloud service providers, which we call machine learning clouds. Machine learning clouds hold the promise of hiding all the sophistication of running large-scale machine learning: Instead of specifying how to run a machine learning task, users only specify what machine learning task to run and the cloud figures out the rest. Raising the level of abstraction, however, rarely comes free - a performance penalty is possible. How good, then, are current machine learning clouds on real-world machine learning workloads? We study this question with a focus on binary classication problems. We present mlbench, a novel benchmark constructed by harvesting datasets from Kaggle competitions. We then compare the performance of the top winning code available from Kaggle with that of running machine learning clouds from both Azure and Amazon on mlbench. Our comparative study reveals the strength and weakness of existing machine learning clouds and points out potential future directions for improvement.
This paper presents the results and conclusions of our participation in the Clickbait Challenge 2017 on automatic clickbait detection in social media. We first describe linguistically-infused neural network models and identify informative representations to predict the level of clickbaiting present in Twitter posts. Our models allow to answer the question not only whether a post is a clickbait or not, but to what extent it is a clickbait post e.g., not at all, slightly, considerably, or heavily clickbaity using a score ranging from 0 to 1. We evaluate the predictive power of models trained on varied text and image representations extracted from tweets. Our best performing model that relies on the tweet text and linguistic markers of biased language extracted from the tweet and the corresponding page yields mean squared error (MSE) of 0.04, mean absolute error (MAE) of 0.16 and R2 of 0.43 on the held-out test data. For the binary classification setup (clickbait vs. non-clickbait), our model achieved F1 score of 0.69. We have not found that image representations combined with text yield significant performance improvement yet. Nevertheless, this work is the first to present preliminary analysis of objects extracted using Google Tensorflow object detection API from images in clickbait vs. non-clickbait Twitter posts. Finally, we outline several steps to improve model performance as a part of the future work.
Binary classification is highly used in credit scoring in the estimation of probability of default. The validation of such predictive models is based both on rank ability, and also on calibration (i.e. how accurately the probabilities output by the model map to the observed probabilities). In this study we cover the current best practices regarding calibration for binary classification, and explore how different approaches yield different results on real world credit scoring data. The limitations of evaluating credit scoring models using only rank ability metrics are explored. A benchmark is run on 18 real world datasets, and results compared. The calibration techniques used are Platt Scaling and Isotonic Regression. Also, different machine learning models are used: Logistic Regression, Random Forest Classifiers, and Gradient Boosting Classifiers. Results show that when the dataset is treated as a time series, the use of re-calibration with Isotonic Regression is able to improve the long term calibration better than the alternative methods. Using re-calibration, the non-parametric models are able to outperform the Logistic Regression on Brier Score Loss.
We consider neural network training, in applications in which there are many possible classes, but at test-time, the task is a binary classification task of determining whether the given example belongs to a specific class, where the class of interest can be different each time the classifier is applied. For instance, this is the case for real-time image search. We define the Single Logit Classification (SLC) task: training the network so that at test-time, it would be possible to accurately identify whether the example belongs to a given class in a computationally efficient manner, based only on the output logit for this class. We propose a natural principle, the Principle of Logit Separation, as a guideline for choosing and designing losses suitable for the SLC. We show that the cross-entropy loss function is not aligned with the Principle of Logit Separation. In contrast, there are known loss functions, as well as novel batch loss functions that we propose, which are aligned with this principle. In total, we study seven loss functions. Our experiments show that indeed in almost all cases, losses that are aligned with the Principle of Logit Separation obtain at least 20% relative accuracy improvement in the SLC task compared to losses that are not aligned with it, and sometimes considerably more. Furthermore, we show that fast SLC does not cause any drop in binary classification accuracy, compared to standard classification in which all logits are computed, and yields a speedup which grows with the number of classes. For instance, we demonstrate a 10x speedup when the number of classes is 400,000. Tensorflow code for optimizing the new batch losses is publicly available at https://github.com/cruvadom/Logit Separation.
Background: Statistical mechanics results (Dauphin et al. (2014); Choromanska et al. (2015)) suggest that local minima with high error are exponentially rare in high dimensions. However, to prove low error guarantees for Multilayer Neural Networks (MNNs), previous works so far required either a heavily modified MNN model or training method, strong assumptions on the labels (e.g., "near" linear separability), or an unrealistic hidden layer with $\Omega\left(N\right)$ units. Results: We examine a MNN with one hidden layer of piecewise linear units, a single output, and a quadratic loss. We prove that, with high probability in the limit of $N\rightarrow\infty$ datapoints, the volume of differentiable regions of the empiric loss containing sub-optimal differentiable local minima is exponentially vanishing in comparison with the same volume of global minima, given standard normal input of dimension $d_{0}=\tilde{\Omega}\left(\sqrt{N}\right)$, and a more realistic number of $d_{1}=\tilde{\Omega}\left(N/d_{0}\right)$ hidden units. We demonstrate our results numerically: for example, $0\%$ binary classification training error on CIFAR with only $N/d_{0}\approx 16$ hidden neurons.
For Time-Domain Global Similarity (TDGS) method, which transforms the data cleaning problem into a binary classification problem about the physical similarity between channels, directly adopting common performance measures could only guarantee the performance for physical similarity. Nevertheless, practical data cleaning tasks have preferences for the correctness of original data sequences. To obtain the general expressions of performance measures based on the preferences of tasks, the mapping relations between performance of TDGS method about physical similarity and correctness of data sequences are investigated by probability theory in this paper. Performance measures for TDGS method in several common data cleaning tasks are set. Cases when these preference-based performance measures could be simplified are introduced.
We propose a new splitting criterion for a meta-learning approach to multiclass classifier design that adaptively merges the classes into a tree-structured hierarchy of increasingly difficult binary classification problems. The classification tree is constructed from empirical estimates of the Henze-Penrose bounds on the pairwise Bayes misclassification rates that rank the binary subproblems in terms of difficulty of classification. The proposed empirical estimates of the Bayes error rate are computed from the minimal spanning tree (MST) of the samples from each pair of classes. Moreover, a meta-learning technique is presented for quantifying the one-vs-rest Bayes error rate for each individual class from a single MST on the entire dataset. Extensive simulations on benchmark datasets show that the proposed hierarchical method can often be learned much faster than competing methods, while achieving competitive accuracy.
This paper introduces a novel real-time Fuzzy Supervised Learning with Binary Meta-Feature (FSL-BM) for big data classification task. The study of real-time algorithms addresses several major concerns, which are namely: accuracy, memory consumption, and ability to stretch assumptions and time complexity. Attaining a fast computational model providing fuzzy logic and supervised learning is one of the main challenges in the machine learning. In this research paper, we present FSL-BM algorithm as an efficient solution of supervised learning with fuzzy logic processing using binary meta-feature representation using Hamming Distance and Hash function to relax assumptions. While many studies focused on reducing time complexity and increasing accuracy during the last decade, the novel contribution of this proposed solution comes through integration of Hamming Distance, Hash function, binary meta-features, binary classification to provide real time supervised method. Hash Tables (HT) component gives a fast access to existing indices; and therefore, the generation of new indices in a constant time complexity, which supersedes existing fuzzy supervised algorithms with better or comparable results. To summarize, the main contribution of this technique for real-time Fuzzy Supervised Learning is to represent hypothesis through binary input as meta-feature space and creating the Fuzzy Supervised Hash table to train and validate model.
Apnea, bradycardia and desaturation (ABD) events often precede life-threatening events including sepsis in newborn babies. Here, we explore machine learning for detection of ABD events as a binary classification problem. We investigate the use of a large neural network to achieve a good detection performance. To be user friendly, the chosen neural network does not require a high level of parameter tuning. Furthermore, a limited amount of training data is available and the training dataset is unbalanced. Comparing with two widely used state-of-the-art machine learning algorithms, the large neural network is found to be efficient. Even with a limited and unbalanced training data, the large neural network provides a detection performance level that is feasible to use in clinical care.
Stereo depth estimation is error-prone; hence, effective error detection methods are desirable. Most such existing methods depend on characteristics of the stereo matching cost curve, making them unduly dependent on functional details of the matching algorithm. As a remedy, we propose a novel error detection approach based solely on the input image and its depth map. Our assumption is that, entropy of any point on an image will be significantly higher than the entropy of its corresponding point on the image's depth map. In this paper, we propose a confidence measure, Entropy-Difference (ED) for stereo depth estimates and a binary classification method to identify incorrect depths. Experiments on the Middlebury dataset show the effectiveness of our method. Our proposed stereo confidence measure outperforms 17 existing measures in all aspects except occlusion detection. Established metrics such as precision, accuracy, recall, and area-under-curve are used to demonstrate the effectiveness of our method.
Towards the vision of translating code that implements an algorithm from one programming language into another, this paper proposes an approach for automated program classification using bilateral tree-based convolutional neural networks (BiTBCNNs). It is layered on top of two tree-based convolutional neural networks (TBCNNs), each of which recognizes the algorithm of code written in an individual programming language. The combination layer of the networks recognizes the similarities and differences among code in different programming languages. The BiTBCNNs are trained using the source code in different languages but known to implement the same algorithms and/or functionalities. For a preliminary evaluation, we use 3591 Java and 3534 C++ code snippets from 6 algorithms we crawled systematically from GitHub. We obtained over 90% accuracy in the cross-language binary classification task to tell whether any given two code snippets implement the same algorithm. Also, for the algorithm classification task, i.e., to predict which one of the six algorithm labels is implemented by an arbitrary C++ code snippet, we achieved over 80% precision.
Data sets for identifying Alzheimer's disease (AD) are often relatively sparse, which limits their ability to train generalizable models. Here, we augment such a data set, DementiaBank, with each of two normative data sets, the Wisconsin Longitudinal Study and Talk2Me, each of which employs a speech-based picture-description assessment. Through minority class oversampling with ADASYN, we outperform state-of-the-art results in binary classification of people with and without AD in DementiaBank. This work highlights the effectiveness of combining sparse and difficult-to-acquire patient data with relatively large and easily accessible normative datasets.
Though suicide is a major public health problem in the US, machine learning methods are not commonly used to predict an individual's risk of attempting/committing suicide. In the present work, starting with an anonymized collection of electronic health records for 522,056 unique, California-resident adolescents, we develop neural network models to predict suicide attempts. We frame the problem as a binary classification problem in which we use a patient's data from 2006-2009 to predict either the presence (1) or absence (0) of a suicide attempt in 2010. After addressing issues such as severely imbalanced classes and the variable length of a patient's history, we build neural networks with depths varying from two to eight hidden layers. For test set observations where we have at least five ED/hospital visits' worth of data on a patient, our depth-4 model achieves a sensitivity of 0.703, specificity of 0.980, and AUC of 0.958.