A technique named Feature Learning from Image Markers (FLIM) was recently proposed to estimate convolutional filters, with no backpropagation, from strokes drawn by a user on very few images (e.g., 1-3) per class, and demonstrated for coconut-tree image classification. This paper extends FLIM for fully connected layers and demonstrates it on different image classification problems. The work evaluates marker selection from multiple users and the impact of adding a fully connected layer. The results show that FLIM-based convolutional neural networks can outperform the same architecture trained from scratch by backpropagation.
This research explores Cost-Sensitive Learning (CSL) in the fraud detection domain to decrease the fraud class's incorrect predictions and increase its accuracy. Notably, we concentrate on shill bidding fraud that is challenging to detect because the behavior of shill and legitimate bidders are similar. We investigate CSL within the Semi-Supervised Classification (SSC) framework to address the scarcity of labeled fraud data. Our paper is the first attempt to integrate CSL with SSC for fraud detection. We adopt a meta-CSL approach to manage the costs of misclassification errors, while SSC algorithms are trained with imbalanced data. Using an actual shill bidding dataset, we assess the performance of several hybrid models of CSL and SSC and then compare their misclassification error and accuracy rates statistically. The most efficient CSL+SSC model was able to detect 99% of fraudsters and with the lowest total cost.
Cluster-and-aggregate techniques such as Vector of Locally Aggregated Descriptors (VLAD), and their end-to-end discriminatively trained equivalents like NetVLAD have recently been popular for video classification and action recognition tasks. These techniques operate by assigning video frames to clusters and then representing the video by aggregating residuals of frames with respect to the mean of each cluster. Since some clusters may see very little video-specific data, these features can be noisy. In this paper, we propose a new cluster-and-aggregate method which we call smoothed Gaussian mixture model (SGMM), and its end-to-end discriminatively trained equivalent, which we call deep smoothed Gaussian mixture model (DSGMM). SGMM represents each video by the parameters of a Gaussian mixture model (GMM) trained for that video. Low-count clusters are addressed by smoothing the video-specific estimates with a universal background model (UBM) trained on a large number of videos. The primary benefit of SGMM over VLAD is smoothing which makes it less sensitive to small number of training samples. We show, through extensive experiments on the YouTube-8M classification task, that SGMM/DSGMM is consistently better than VLAD/NetVLAD by a small but statistically significant margin. We also show results using a dataset created at LinkedIn to predict if a member will watch an uploaded video.
The problem of class imbalance is extensive for focusing on numerous applications in the real world. In such a situation, nearly all of the examples are labeled as one class called majority class, while far fewer examples are labeled as the other class usually, the more important class is called minority. Over the last few years, several types of research have been carried out on the issue of class imbalance, including data sampling, cost-sensitive analysis, Genetic Programming based models, bagging, boosting, etc. Nevertheless, in this survey paper, we enlisted the 24 related studies in the years 2003, 2008, 2010, 2012 and 2014 to 2019, focusing on the architecture of single, hybrid, and ensemble method design to understand the current status of improving classification output in machine learning techniques to fix problems with class imbalances. This survey paper also includes a statistical analysis of the classification algorithms under various methods and several other experimental conditions, as well as datasets used in different research papers.
With COVID-19 cases rising rapidly, deep learning has emerged as a promising diagnosis technique. However, identifying the most accurate models to characterize COVID-19 patients is challenging because comparing results obtained with different types of data and acquisition processes is non-trivial. In this paper we designed, evaluated, and compared the performance of 20 convolutional neutral networks in classifying patients as COVID-19 positive, healthy, or suffering from other pulmonary lung infections based on Chest CT scans, serving as the first to consider the EfficientNet family for COVID-19 diagnosis and employ intermediate activation maps for visualizing model performance. All models are trained and evaluated in Python using 4173 Chest CT images from the dataset entitled "A COVID multiclass dataset of CT scans," with 2168, 758, and 1247 images of patients that are COVID-19 positive, healthy, or suffering from other pulmonary infections, respectively. EfficientNet-B5 was identified as the best model with an F1 score of 0.9769+/-0.0046, accuracy of 0.9759+/-0.0048, sensitivity of 0.9788+/-0.0055, specificity of 0.9730+/-0.0057, and precision of 0.9751 +/- 0.0051. On an alternate 2-class dataset, EfficientNetB5 obtained an accuracy of 0.9845+/-0.0109, F1 score of 0.9599+/-0.0251, sensitivity of 0.9682+/-0.0099, specificity of 0.9883+/-0.0150, and precision of 0.9526 +/- 0.0523. Intermediate activation maps and Gradient-weighted Class Activation Mappings offered human-interpretable evidence of the model's perception of ground-class opacities and consolidations, hinting towards a promising use-case of artificial intelligence-assisted radiology tools. With a prediction speed of under 0.1 seconds on GPUs and 0.5 seconds on CPUs, our proposed model offers a rapid, scalable, and accurate diagnostic for COVID-19.
New binary classification tests are often evaluated relative to a pre-established test. For example, rapid Antigen tests for the detection of SARS-CoV-2 are assessed relative to more established PCR tests. In this paper, I argue that the new test can be described as producing ambiguous information when the pre-established is imperfect. This allows for a phenomenon called dilation -- an extreme form of non-informativeness. As an example, I present hypothetical test data satisfying the WHO's minimum quality requirement for rapid Antigen tests which leads to dilation. The ambiguity in the information arises from a missing data problem due to imperfection of the established test: the joint distribution of true infection and test results is not observed. Using results from Copula theory, I construct the (usually non-singleton) set of all these possible joint distributions, which allows me to assess the new test's informativeness. This analysis leads to a simple sufficient condition to make sure that a new test is not a dilation. I illustrate my approach with applications to data from three COVID-19 related tests. Two rapid Antigen tests satisfy my sufficient condition easily and are therefore informative. However, less accurate procedures, like chest CT scans, may exhibit dilation.
This paper illustrates the details description of technical text classification system and its results that developed as a part of participation in the shared task TechDofication 2020. The shared task consists of two sub-tasks: (i) first task identify the coarse-grained technical domain of given text in a specified language and (ii) the second task classify a text of computer science domain into fine-grained sub-domains. A classification system (called 'TechTexC') is developed to perform the classification task using three techniques: convolution neural network (CNN), bidirectional long short term memory (BiLSTM) network, and combined CNN with BiLSTM. Results show that CNN with BiLSTM model outperforms the other techniques concerning task-1 of sub-tasks (a, b, c and g) and task-2a. This combined model obtained f1 scores of 82.63 (sub-task a), 81.95 (sub-task b), 82.39 (sub-task c), 84.37 (sub-task g), and 67.44 (task-2a) on the development dataset. Moreover, in the case of test set, the combined CNN with BiLSTM approach achieved that higher accuracy for the subtasks 1a (70.76%), 1b (79.97%), 1c (65.45%), 1g (49.23%) and 2a (70.14%).
Fine-grained visual classification (FGVC) aims to distinguish the sub-classes of the same category and its essential solution is to mine the subtle and discriminative regions. Convolution neural networks (CNNs), which employ the cross entropy loss (CE-loss) as the loss function, show poor performance since the model can only learn the most discriminative part and ignore other meaningful regions. Some existing works try to solve this problem by mining more discriminative regions by some detection techniques or attention mechanisms. However, most of them will meet the background noise problem when trying to find more discriminative regions. In this paper, we address it in a knowledge transfer learning manner. Multiple models are trained one by one, and all previously trained models are regarded as teacher models to supervise the training of the current one. Specifically, a orthogonal loss (OR-loss) is proposed to encourage the network to find diverse and meaningful regions. In addition, the first model is trained with only CE-Loss. Finally, all models' outputs with complementary knowledge are combined together for the final prediction result. We demonstrate the superiority of the proposed method and obtain state-of-the-art (SOTA) performances on three popular FGVC datasets.
Convolutional Networks have dominated the field of computer vision for the last ten years, exhibiting extremely powerful feature extraction capabilities and outstanding classification performance. The main strategy to prolong this trend relies on further upscaling networks in size. However, costs increase rapidly while performance improvements may be marginal. We hypothesise that adding heterogeneous sources of information may be more cost-effective to a CNN than building a bigger network. In this paper, an ensemble method is proposed for accurate image classification, fusing automatically detected features through Convolutional Neural Network architectures with a set of manually defined statistical indicators. Through a combination of the predictions of a CNN and a secondary classifier trained on statistical features, better classification performance can be cheaply achieved. We test multiple learning algorithms and CNN architectures on a diverse number of datasets to validate our proposal, making public all our code and data via GitHub. According to our results, the inclusion of additional indicators and an ensemble classification approach helps to increase the performance in 8 of 9 datasets, with a remarkable increase of more than 10% precision in two of them.
This manuscript presents a new framework towards automated digital documentation and progress reporting of mechanical pipes in building construction projects, using smartphones. New methods were proposed to optimize video frame rate to achieve a desired image overlap; define metric scale for 3D reconstruction; extract pipes from point clouds; and classify pipes according to their planned bill of quantity radii. The effectiveness of the proposed methods in both laboratory (six pipes) and construction site (58 pipes) conditions was evaluated. It was observed that the proposed metric scale definition achieved sub-millimeter pipe radius estimation accuracy. Both laboratory and field experiments revealed that increasing the defined image overlap improved point cloud quality, pipe classification quality, and pipe radius/length estimation. Overall, it was found possible to achieve pipe classification F-measure, radius estimation accuracy, and length estimation percent error of 96.4%, 5.4mm, and 5.0%, respectively, on construction sites using at least 95% image overlap.
Hyperspectral image classification (HIC) is an important but challenging task, and a problem that limits the algorithmic development in this field is that the ground truths of hyperspectral images (HSIs) are extremely hard to obtain. Recently a handful of HIC methods are developed based on the graph convolution networks (GCNs), which effectively relieves the scarcity of labeled data for deep learning based HIC methods. To further lift the classification performance, in this work we propose a graph convolution network (GCN) based framework for HSI classification that uses two clustering operations to better exploit multi-hop node correlations and also effectively reduce graph size. In particular, we first cluster the pixels with similar spectral features into a superpixel and build the graph based on the superpixels of the input HSI. Then instead of performing convolution over this superpixel graph, we further partition it into several sub-graphs by pruning the edges with weak weights, so as to strengthen the correlations of nodes with high similarity. This second round of clustering also further reduces the graph size, thus reducing the computation burden of graph convolution. Experimental results on three widely used benchmark datasets well prove the effectiveness of our proposed framework.
Effective analysis of EEG signals for potential clinical applications remains a challenging task. So far, the analysis and conditioning of EEG have largely remained sex-neutral. This paper employs a machine learning approach to explore the evidence of sex effects on EEG signals, and confirms the generality of these effects by achieving successful sex prediction of resting-state EEG signals. We have found that the brain connectivity represented by the coherence between certain sensor channels are good predictors of sex.
In this paper, we propose a novel method named Biomedical Confident Itemsets Explanation (BioCIE), aiming at post-hoc explanation of black-box machine learning models for biomedical text classification. Using sources of domain knowledge and a confident itemset mining method, BioCIE discretizes the decision space of a black-box into smaller subspaces and extracts semantic relationships between the input text and class labels in different subspaces. Confident itemsets discover how biomedical concepts are related to class labels in the black-box's decision space. BioCIE uses the itemsets to approximate the black-box's behavior for individual predictions. Optimizing fidelity, interpretability, and coverage measures, BioCIE produces class-wise explanations that represent decision boundaries of the black-box. Results of evaluations on various biomedical text classification tasks and black-box models demonstrated that BioCIE can outperform perturbation-based and decision set methods in terms of producing concise, accurate, and interpretable explanations. BioCIE improved the fidelity of instance-wise and class-wise explanations by 11.6% and 7.5%, respectively. It also improved the interpretability of explanations by 8%. BioCIE can be effectively used to explain how a black-box biomedical text classification model semantically relates input texts to class labels. The source code and supplementary material are available at https://github.com/mmoradi-iut/BioCIE.
Tree based ensembles such as Breiman's random forest (RF) and Gradient Boosted Trees (GBT) can be interpreted as implicit kernel generators, where the ensuing proximity matrix represents the data-driven tree ensemble kernel. Kernel perspective on the RF has been used to develop a principled framework for theoretical investigation of its statistical properties. Recently, it has been shown that the kernel interpretation is germane to other tree-based ensembles e.g. GBTs. However, practical utility of the links between kernels and the tree ensembles has not been widely explored and systematically evaluated. Focus of our work is investigation of the interplay between kernel methods and the tree based ensembles including the RF and GBT. We elucidate the performance and properties of the RF and GBT based kernels in a comprehensive simulation study comprising of continuous and binary targets. We show that for continuous targets, the RF/GBT kernels are competitive to their respective ensembles in higher dimensional scenarios, particularly in cases with larger number of noisy features. For the binary target, the RF/GBT kernels and their respective ensembles exhibit comparable performance. We provide the results from real life data sets for regression and classification to show how these insights may be leveraged in practice. Overall, our results support the tree ensemble based kernels as a valuable addition to the practitioner's toolbox. Finally, we discuss extensions of the tree ensemble based kernels for survival targets, interpretable prototype and landmarking classification and regression. We outline future line of research for kernels furnished by Bayesian counterparts of the frequentist tree ensembles.
We consider the problem of classifying a map using a team of communicating robots. It is assumed that all robots have localized visual sensing capabilities and can exchange their information with neighboring robots. Using a graph decomposition technique, we proposed an offline learning structure that makes every robot capable of communicating with and fusing information from its neighbors to plan its next move towards the most informative parts of the environment for map classification purposes. The main idea is to decompose a given undirected graph into a union of directed star graphs and train robots w.r.t a bounded number of star graphs. This will significantly reduce the computational cost of offline training and makes learning scalable (independent of the number of robots). Our approach is particularly useful for fast map classification in large environments using a large number of communicating robots. We validate the usefulness of our proposed methodology through extensive simulations.
Time series classification problems exist in many fields and have been explored for a couple of decades. However, they still remain challenging, and their solutions need to be further improved for real-world applications in terms of both accuracy and efficiency. In this paper, we propose a hybrid neural architecture, called Self-Attentive Recurrent Convolutional Networks (SARCoN), to learn multi-faceted representations for univariate time series. SARCoN is the synthesis of long short-term memory networks with self-attentive mechanisms and Fully Convolutional Networks, which work in parallel to learn the representations of univariate time series from different perspectives. The component modules of the proposed architecture are trained jointly in an end-to-end manner and they classify the input time series in a cooperative way. Due to its domain-agnostic nature, SARCoN is able to generalize a diversity of domain tasks. Our experimental results show that, compared to the state-of-the-art approaches for time series classification, the proposed architecture can achieve remarkable improvements for a set of univariate time series benchmarks from the UCR repository. Moreover, the self-attention and the global average pooling in the proposed architecture enable visible interpretability by facilitating the identification of the contribution regions of the original time series. An overall analysis confirms that multi-faceted representations of time series aid in capturing deep temporal corrections within complex time series, which is essential for the improvement of time series classification performance. Our work provides a novel angle that deepens the understanding of time series classification, qualifying our proposed model as an ideal choice for real-world applications.
We consider the sample complexity of learning with adversarial robustness. Most prior theoretical results for this problem have considered a setting where different classes in the data are close together or overlapping. Motivated by some real applications, we consider, in contrast, the well-separated case where there exists a classifier with perfect accuracy and robustness, and show that the sample complexity narrates an entirely different story. Specifically, for linear classifiers, we show a large class of well-separated distributions where the expected robust loss of any algorithm is at least $\Omega(\frac{d}{n})$, whereas the max margin algorithm has expected standard loss $O(\frac{1}{n})$. This shows a gap in the standard and robust losses that cannot be obtained via prior techniques. Additionally, we present an algorithm that, given an instance where the robustness radius is much smaller than the gap between the classes, gives a solution with expected robust loss is $O(\frac{1}{n})$. This shows that for very well-separated data, convergence rates of $O(\frac{1}{n})$ are achievable, which is not the case otherwise. Our results apply to robustness measured in any $\ell_p$ norm with $p > 1$ (including $p = \infty$).
Graph neural networks (GNNs) have been investigated for potential applicability in multiple fields that employ graph data. However, there are no standard training settings to ensure fair comparisons among new methods, including different model architectures and data augmentation techniques. We introduce a standard, reproducible benchmark to which the same training settings can be applied for node classification. For this benchmark, we constructed 9 datasets, including both small- and medium-scale datasets from different fields, and 7 different models. We design a k-fold model assessment strategy for small datasets and a standard set of model training procedures for all datasets, enabling a standard experimental pipeline for GNNs to help ensure fair model architecture comparisons. We use node2vec and Laplacian eigenvectors to perform data augmentation to investigate how input features affect the performance of the models. We find topological information is important for node classification tasks. Increasing the number of model layers does not improve the performance except on the PATTERN and CLUSTER datasets, in which the graphs are not connected. Data augmentation is highly useful, especially using node2vec in the baseline, resulting in a substantial baseline performance improvement.
Algorithms for noiseless collaborative PAC learning have been analyzed and optimized in recent years with respect to sample complexity. In this paper, we study collaborative PAC learning with the goal of reducing communication cost at essentially no penalty to the sample complexity. We develop communication efficient collaborative PAC learning algorithms using distributed boosting. We then consider the communication cost of collaborative learning in the presence of classification noise. As an intermediate step, we show how collaborative PAC learning algorithms can be adapted to handle classification noise. With this insight, we develop communication efficient algorithms for collaborative PAC learning robust to classification noise.
Small business classification is a difficult and important task within many applications, including customer segmentation. Training on small business names introduces gender and geographic origin biases. A model for predicting one of 66 business types based only upon the business name was developed in this work (top-1 f1-score = 60.2%). Two approaches to removing the bias from this model are explored: replacing given names with a placeholder token, and augmenting the training data with gender-swapped examples. The results for these approaches is reported, and the bias in the model was reduced by hiding given names from the model. However, bias reduction was accomplished at the expense of classification performance (top-1 f1-score = 56.6%). Augmentation of the training data with gender-swapping samples proved less effective at bias reduction than the name hiding approach on the evaluated dataset.