Character-level convolutional neural networks (char-CNN) require no knowledge of the semantic or syntactic structure of the language they classify. This property simplifies its implementation but reduces its classification accuracy. Increasing the depth of char-CNN architectures does not result in breakthrough accuracy improvements. Research has not established which char-CNN architectures are optimal for text classification tasks. Manually designing and training char-CNNs is an iterative and time-consuming process that requires expert domain knowledge. Evolutionary deep learning (EDL) techniques, including surrogate-based versions, have demonstrated success in automatically searching for performant CNN architectures for image analysis tasks. Researchers have not applied EDL techniques to search the architecture space of char-CNNs for text classification tasks. This article demonstrates the first work in evolving char-CNN architectures using a novel EDL algorithm based on genetic programming, an indirect encoding and surrogate models, to search for performant char-CNN architectures automatically. The algorithm is evaluated on eight text classification datasets and benchmarked against five manually designed CNN architecture and one long short-term memory (LSTM) architecture. Experiment results indicate that the algorithm can evolve architectures that outperform the LSTM in terms of classification accuracy and five of the manually designed CNN architectures in terms of classification accuracy and parameter count.
We present a machine learning based COVID-19 cough classifier which can discriminate COVID-19 positive coughs from both COVID-19 negative and healthy coughs recorded on a smartphone. This type of screening is non-contact, easy to apply, and can reduce the workload in testing centres as well as limit transmission by recommending early self-isolation to those who have a cough suggestive of COVID-19. The datasets used in this study include subjects from all six continents and contain both forced and natural coughs, indicating that the approach is widely applicable. The publicly available Coswara dataset contains 92 COVID-19 positive and 1079 healthy subjects, while the second smaller dataset was collected mostly in South Africa and contains 18 COVID-19 positive and 26 COVID-19 negative subjects who have undergone a SARS-CoV laboratory test. Both datasets indicate that COVID-19 positive coughs are 15\%-20\% shorter than non-COVID coughs. Dataset skew was addressed by applying the synthetic minority oversampling technique (SMOTE). A leave-$p$-out cross-validation scheme was used to train and evaluate seven machine learning classifiers: LR, KNN, SVM, MLP, CNN, LSTM and Resnet50. Our results show that although all classifiers were able to identify COVID-19 coughs, the best performance was exhibited by the Resnet50 classifier, which was best able to discriminate between the COVID-19 positive and the healthy coughs with an area under the ROC curve (AUC) of 0.98. An LSTM classifier was best able to discriminate between the COVID-19 positive and COVID-19 negative coughs, with an AUC of 0.94 after selecting the best 13 features from a sequential forward selection (SFS). Since this type of cough audio classification is cost-effective and easy to deploy, it is potentially a useful and viable means of non-contact COVID-19 screening.
By transferring knowledge learned from seen/previous tasks, meta learning aims to generalize well to unseen/future tasks. Existing meta-learning approaches have shown promising empirical performance on various multiclass classification problems, but few provide theoretical analysis on the classifiers' generalization ability on future tasks. In this paper, under the assumption that all classification tasks are sampled from the same meta-distribution, we leverage margin theory and statistical learning theory to establish three margin-based transfer bounds for meta-learning based multiclass classification (MLMC). These bounds reveal that the expected error of a given classification algorithm for a future task can be estimated with the average empirical error on a finite number of previous tasks, uniformly over a class of preprocessing feature maps/deep neural networks (i.e. deep feature embeddings). To validate these bounds, instead of the commonly-used cross-entropy loss, a multi-margin loss is employed to train a number of representative MLMC models. Experiments on three benchmarks show that these margin-based models still achieve competitive performance, validating the practical value of our margin-based theoretical analysis.
Committee-based models (ensembles or cascades) construct models by combining existing pre-trained ones. While ensembles and cascades are well-known techniques that were proposed before deep learning, they are not considered a core building block of deep model architectures and are rarely compared to in recent literature on developing efficient models. In this work, we go back to basics and conduct a comprehensive analysis of the efficiency of committee-based models. We find that even the most simplistic method for building committees from existing, independently pre-trained models can match or exceed the accuracy of state-of-the-art models while being drastically more efficient. These simple committee-based models also outperform sophisticated neural architecture search methods (e.g., BigNAS). These findings hold true for several tasks, including image classification, video classification, and semantic segmentation, and various architecture families, such as ViT, EfficientNet, ResNet, MobileNetV2, and X3D. Our results show that an EfficientNet cascade can achieve a 5.4x speedup over B7 and a ViT cascade can achieve a 2.3x speedup over ViT-L-384 while being equally accurate.
Physical human-robot interactions (pHRI) are less efficient and communicative than human-human interactions, and a key reason is a lack of informative sense of touch in robotic systems. Interpreting human touch gestures is a nuanced, challenging task with extreme gaps between human and robot capability. Among prior works that demonstrate human touch recognition capability, differences in sensors, gesture classes, feature sets, and classification algorithms yield a conglomerate of non-transferable results and a glaring lack of a standard. To address this gap, this work presents 1) four proposed touch gesture classes that cover an important subset of the gesture characteristics identified in the literature, 2) the collection of an extensive force dataset on a common pHRI robotic arm with only its internal wrist force-torque sensor, and 3) an exhaustive performance comparison of combinations of feature sets and classification algorithms on this dataset. We demonstrate high classification accuracies among our proposed gesture definitions on a test set, emphasizing that neural net-work classifiers on the raw data outperform other combinations of feature sets and algorithms. The accompanying video is here: https://youtu.be/gJPVImNKU68
A long-standing issue with deep learning is the need for large and consistently labeled datasets. Although the current research in semi-supervised learning can decrease the required amount of annotated data by a factor of 10 or even more, this line of research still uses distinct classes like cats and dogs. However, in the real-world we often encounter problems where different experts have different opinions, thus producing fuzzy labels. We propose a novel framework for handling semi-supervised classifications of such fuzzy labels. Our framework is based on the idea of overclustering to detect substructures in these fuzzy labels. We propose a novel loss to improve the overclustering capability of our framework and show on the common image classification dataset STL-10 that it is faster and has better overclustering performance than previous work. On a real-world plankton dataset, we illustrate the benefit of overclustering for fuzzy labels and show that we beat previous state-of-the-art semisupervised methods. Moreover, we acquire 5 to 10% more consistent predictions of substructures.
Advances in deep learning have resulted in state-of-the-art performance for many audio classification tasks but, unlike humans, these systems traditionally require large amounts of data to make accurate predictions. Not every person or organization has access to those resources, and the organizations that do, like our field at large, do not reflect the demographics of our country. Enabling people to use machine learning without significant resource hurdles is important, because machine learning is an increasingly useful tool for solving problems, and can solve a broader set of problems when put in the hands of a broader set of people. Few-shot learning is a type of machine learning designed to enable the model to generalize to new classes with very few examples. In this research, we address two audio classification tasks (speaker identification and activity classification) with the Prototypical Network few-shot learning algorithm, and assess performance of various encoder architectures. Our encoders include recurrent neural networks, as well as one- and two-dimensional convolutional neural networks. We evaluate our model for speaker identification on the VoxCeleb dataset and ICSI Meeting Corpus, obtaining 5-shot 5-way accuracies of 93.5% and 54.0%, respectively. We also evaluate for activity classification from audio using few-shot subsets of the Kinetics~600 dataset and AudioSet, both drawn from Youtube videos, obtaining 51.5% and 35.2% accuracy, respectively.
In this paper, we leverage predictive uncertainty of deep neural networks to answer challenging questions material scientists usually encounter in machine learning based materials applications workflows. First, we show that by leveraging predictive uncertainty, a user can determine the required training data set size necessary to achieve a certain classification accuracy. Next, we propose uncertainty guided decision referral to detect and refrain from making decisions on confusing samples. Finally, we show that predictive uncertainty can also be used to detect out-of-distribution test samples. We find that this scheme is accurate enough to detect a wide range of real-world shifts in data, e.g., changes in the image acquisition conditions or changes in the synthesis conditions. Using microstructure information from scanning electron microscope (SEM) images as an example use case, we show that leveraging uncertainty-aware deep learning can significantly improve the performance and dependability of classification models.
Optical colonoscopy is an essential diagnostic and prognostic tool for many gastrointestinal diseases, including cancer screening and staging, intestinal bleeding, diarrhea, abdominal symptom evaluation, and inflammatory bowel disease assessment. Automated assessment of colonoscopy is of interest considering the subjectivity present in qualitative human interpretations of colonoscopy findings. Localization of the camera is essential to interpreting the meaning and context of findings for diseases evaluated by colonoscopy. In this study, we propose a camera localization system to estimate the relative location of the camera and classify the colon into anatomical segments. The camera localization system begins with non-informative frame detection and removal. Then a self-training end-to-end convolutional neural network is built to estimate the camera motion, where several strategies are proposed to improve its robustness and generalization on endoscopic videos. Using the estimated camera motion a camera trajectory can be derived and a relative location index calculated. Based on the estimated location index, anatomical colon segment classification is performed by constructing a colon template. The proposed motion estimation algorithm was evaluated on an external dataset containing the ground truth for camera pose. The experimental results show that the performance of the proposed method is superior to other published methods. The relative location index estimation and anatomical region classification were further validated using colonoscopy videos collected from routine clinical practice. This validation yielded an average accuracy in classification of 0.754, which is substantially higher than the performances obtained using location indices built from other methods.
Learning to generate a task-aware base learner proves a promising direction to deal with few-shot learning (FSL) problem. Existing methods mainly focus on generating an embedding model utilized with a fixed metric (eg, cosine distance) for nearest neighbour classification or directly generating a linear classier. However, due to the limited discriminative capacity of such a simple metric or classifier, these methods fail to generalize to challenging cases appropriately. To mitigate this problem, we present a novel deep metric meta-generation method that turns to an orthogonal direction, ie, learning to adaptively generate a specific metric for a new FSL task based on the task description (eg, a few labelled samples). In this study, we structure the metric using a three-layer deep attentive network that is flexible enough to produce a discriminative metric for each task. Moreover, different from existing methods that utilize an uni-modal weight distribution conditioned on labelled samples for network generation, the proposed meta-learner establishes a multi-modal weight distribution conditioned on cross-class sample pairs using a tailored variational autoencoder, which can separately capture the specific inter-class discrepancy statistics for each class and jointly embed the statistics for all classes into metric generation. By doing this, the generated metric can be appropriately adapted to a new FSL task with pleasing generalization performance. To demonstrate this, we test the proposed method on four benchmark FSL datasets and gain surprisingly obvious performance improvement over state-of-the-art competitors, especially in the challenging cases, eg, improve the accuracy from 26.14% to 46.69% in the 20-way 1-shot task on miniImageNet, while improve the accuracy from 45.2% to 68.72% in the 5-way 1-shot task on FC100. Code is available: https://github.com/NWPUZhoufei/DAM.
We apply deep-neural-network-based techniques to quantum state classification and reconstruction. We demonstrate high classification accuracies and reconstruction fidelities, even in the presence of noise and with little data. Using optical quantum states as examples, we first demonstrate how convolutional neural networks (CNNs) can successfully classify several types of states distorted by, e.g., additive Gaussian noise or photon loss. We further show that a CNN trained on noisy inputs can learn to identify the most important regions in the data, which potentially can reduce the cost of tomography by guiding adaptive data collection. Secondly, we demonstrate reconstruction of quantum-state density matrices using neural networks that incorporate quantum-physics knowledge. The knowledge is implemented as custom neural-network layers that convert outputs from standard feedforward neural networks to valid descriptions of quantum states. Any standard feed-forward neural-network architecture can be adapted for quantum state tomography (QST) with our method. We present further demonstrations of our proposed [arXiv:2008.03240] QST technique with conditional generative adversarial networks (QST-CGAN). We motivate our choice of a learnable loss function within an adversarial framework by demonstrating that the QST-CGAN outperforms, across a range of scenarios, generative networks trained with standard loss functions. For pure states with additive or convolutional Gaussian noise, the QST-CGAN is able to adapt to the noise and reconstruct the underlying state. The QST-CGAN reconstructs states using up to two orders of magnitude fewer iterative steps than a standard iterative maximum likelihood (iMLE) method. Further, the QST-CGAN can reconstruct both pure and mixed states from two orders of magnitude fewer randomly chosen data points than iMLE.
We are in the era where the Big Data analytics has changed the way of interpreting the various biomedical phenomena, and as the generated data increase, the need for new machine learning methods to handle this evolution grows. An indicative example is the single-cell RNA-seq (scRNA-seq), an emerging DNA sequencing technology with promising capabilities but significant computational challenges due to the large-scaled generated data. Regarding the classification process for scRNA-seq data, an appropriate method is the k Nearest Neighbor (kNN) classifier since it is usually utilized for large-scale prediction tasks due to its simplicity, minimal parameterization, and model-free nature. However, the ultra-high dimensionality that characterizes scRNA-seq impose a computational bottleneck, while prediction power can be affected by the "Curse of Dimensionality". In this work, we proposed the utilization of approximate nearest neighbor search algorithms for the task of kNN classification in scRNA-seq data focusing on a particular methodology tailored for high dimensional data. We argue that even relaxed approximate solutions will not affect the prediction performance significantly. The experimental results confirm the original assumption by offering the potential for broader applicability.
In this paper we reformulate few-shot classification as a reconstruction problem in latent space. The ability of the network to reconstruct a query feature map from support features of a given class predicts membership of the query in that class. We introduce a novel mechanism for few-shot classification by regressing directly from support features to query features in closed form, without introducing any new modules or large-scale learnable parameters. The resulting Feature Map Reconstruction Networks are both more performant and computationally efficient than previous approaches. We demonstrate consistent and substantial accuracy gains on four fine-grained benchmarks with varying neural architectures. Our model is also competitive on the non-fine-grained mini-ImageNet and tiered-ImageNet benchmarks with minimal bells and whistles.
Fairness in visual recognition is becoming a prominent and critical topic of discussion as recognition systems are deployed at scale in the real world. Models trained from data in which target labels are correlated with protected attributes (e.g., gender, race) are known to learn and exploit those correlations. In this work, we introduce a method for training accurate target classifiers while mitigating biases that stem from these correlations. We use GANs to generate realistic-looking images, and perturb these images in the underlying latent space to generate training data that is balanced for each protected attribute. We augment the original dataset with this perturbed generated data, and empirically demonstrate that target classifiers trained on the augmented dataset exhibit a number of both quantitative and qualitative benefits. We conduct a thorough evaluation across multiple target labels and protected attributes in the CelebA dataset, and provide an in-depth analysis and comparison to existing literature in the space.
Stock classification is a challenging task due to high levels of noise and volatility of stocks returns. In this paper we show that using transfer learning can help with this task, by pre-training a model to extract universal features on the full universe of stocks of the S$\&$P500 index and then transferring it to another model to directly learn a trading rule. Transferred models present more than double the risk-adjusted returns than their counterparts trained from zero. In addition, we propose the use of data augmentation on the feature space defined as the output of a pre-trained model (i.e. augmenting the aggregated time-series representation). We compare this augmentation approach with the standard one, i.e. augmenting the time-series in the input space. We show that augmentation methods on the feature space leads to $20\%$ increase in risk-adjusted return compared to a model trained with transfer learning but without augmentation.
Atrial Fibrillation (AF) is among one of the most common types of heart arrhythmia afflicting more than 3 million people in the U.S. alone. AF is estimated to be the cause of death of 1 in 4 individuals. Recent advancements in Artificial Intelligence (AI) algorithms have led to the capability of reliably detecting AF from ECG signals. While these algorithms can accurately detect AF with high precision, the discrete and deterministic classifications mean that these networks are likely to erroneously classify the given ECG signal. This paper proposes a variational autoencoder classifier network that provides an uncertainty estimation of the network's output in addition to reliable classification accuracy. This framework can increase physicians' trust in using AI-based AF detection algorithms by providing them with a confidence score which reflects how uncertain the algorithm is about a case and recommending them to put more attention to the cases with a lower confidence score. The uncertainty is estimated by conducting multiple passes of the input through the network to build a distribution; the mean of the standard deviations is reported as the network's uncertainty. Our proposed network obtains 97.64% accuracy in addition to reporting the uncertainty
This paper presents a novel method for classifying radio frequency (RF) devices from their transmission signals. Given a collection of signals from identical devices, we accurately classify both the distance of the transmission and the specific device identity. We develop a multiple classifier system that accurately discriminates between channels and classifies devices using normalized in-phase and quadrature (IQ) samples. Our network uses residual connections for both distance and device classification, reaching 88.33% accuracy classifying 16 unique devices over 11 different distances and two different times, on a task that was previously unlearnable. Furthermore, we demonstrate the efficacy for pre-training neural networks for massive data domains and subtle classification differences.
As the threats of small drones increase, not only the detection but also the classification of small drones has become important. Many recent studies have applied an approach to utilize the micro-Doppler signature (MDS) for the small drone classification by using frequency modulated continuous wave (FMCW) radars. In this letter, we propose a novel method to extract the MDS images of the small drones with the FMCW radar. Moreover, we propose a light convolutional neural network (CNN) whose structure is straightforward, and the number of parameters is quite small for fast classification. The proposed method contributes to increasing the classification accuracy by improving the quality of MDS images. We classified the small drones with the MDS images extracted by the conventional method and the proposed method through the proposed CNN. The experimental results showed that the total classification accuracy was increased by 10.00 % due to the proposed method. The total classification accuracy was recorded at 97.14 % with the proposed MDS extraction method and the proposed light CNN.
The recent advances in Gravitational-wave astronomy have greatly accelerated the study of Multimessenger astrophysics. There is a need for the development of fast and efficient algorithms to detect non-astrophysical transients and noises due to the rate and scale at which the data is being provided by LIGO and other gravitational wave observatories. These transients and noises can interfere with the study of gravitational waves and binary mergers and induce false positives. Here, we propose the use of deep learning algorithms to detect and classify these transient signals. Traditional statistical methods are not well designed for dealing with temporal signals but supervised deep learning techniques such as RNN-LSTM and deep CNN have proven to be effective for solving problems such as time-series forecasting and time-series classification. We also use unsupervised models such as Total variation, Principal Component Analysis, Support Vector Machine, Wavelet decomposition or Random Forests for feature extraction and noise reduction and then study the results obtained by RNN-LSTM and deep CNN for classifying the transients in low-SNR signals. We compare the results obtained by the combination of various unsupervised models and supervised models. This method can be extended to real-time detection of transients and merger signals using deep-learning optimized GPU's for early prediction and study of various astronomical events. We will also explore and compare other machine learning models such as MLP, Stacked Autoencoder, Random forests, extreme learning machine, Support Vector machine and logistic regression classifier.
With the increased interest in integrating renewable energy sources (RES) such as wind power and solar into the power systems owing to their zero greenhouse gas emissions and the involvement of power converters for integration in grid, the detection, classification and mitigation of power quality events has become indispensable. For employing an appropriate mitigation technique, it is a pre-requisite to correctly classify the various types of disturbances in power quality. This paper, therefore, presents a detailed research reviews on detection and classification of power quality disturbances caused by the integration of renewable energy sources and associated works present in literature till date. Attempts are also made to highlight the current and future issues involved in the detection, classification and mitigation of PQ disturbances. Best efforts have been made to make this paper serve as a full-fledged reference for the future work in this field. A list of 230 research publications on the subject is also appended for quick reference.