Transformers in their common form are inherently limited to operate on whole token sequences rather than on one token at a time. Consequently, their use during online inference on time-series data entails considerable redundancy due to the overlap in successive token sequences. In this work, we propose novel formulations of the Scaled Dot-Product Attention, which enable Transformers to perform efficient online token-by-token inference on a continual input stream. Importantly, our modifications are purely to the order of computations, while the outputs and learned weights are identical to those of the original Transformer Encoder. We validate our Continual Transformer Encoder with experiments on the THUMOS14, TVSeries and GTZAN datasets with remarkable results: Our Continual one- and two-block architectures reduce the floating point operations per prediction by up to 63x and 2.6x, respectively, while retaining predictive performance.
Cervical cancer is the fourth most common cancer in women worldwide. The availability of a robust automated cervical image classification system can augment the clinical care provider's limitation in traditional visual inspection with acetic acid (VIA). However, there are a wide variety of cervical inspection objectives which impact the labeling criteria for criteria-specific prediction model development. Moreover, due to the lack of confirmatory test results and inter-rater labeling variation, many images are left unlabeled. Motivated by these challenges, we propose a self-supervised learning (SSL) based approach to produce a pre-trained cervix model from unlabeled cervical images. The developed model is further fine-tuned to produce criteria-specific classification models with the available labeled images. We demonstrate the effectiveness of the proposed approach using two cervical image datasets. Both datasets are partially labeled and labeling criteria are different. The experimental results show that the SSL-based initialization improves classification performance (Accuracy: 2.5% min) and the inclusion of images from both datasets during SSL further improves the performance (Accuracy: 1.5% min). Further, considering data-sharing restrictions, we experimented with the effectiveness of Federated SSL and find that it can improve performance over the SSL model developed with just its images. This justifies the importance of SSL-based cervix model development. We believe that the present research shows a novel direction in developing criteria-specific custom deep models for cervical image classification by combining images from different sources unlabeled and/or labeled with varying criteria, and addressing image access restrictions.
Hate speech is a major issue in social networks due to the high volume of data generated daily. Recent works demonstrate the usefulness of machine learning (ML) in dealing with the nuances required to distinguish between hateful posts from just sarcasm or offensive language. Many ML solutions for hate speech detection have been proposed by either changing how features are extracted from the text or the classification algorithm employed. However, most works consider only one type of feature extraction and classification algorithm. This work argues that a combination of multiple feature extraction techniques and different classification models is needed. We propose a framework to analyze the relationship between multiple feature extraction and classification techniques to understand how they complement each other. The framework is used to select a subset of complementary techniques to compose a robust multiple classifiers system (MCS) for hate speech detection. The experimental study considering four hate speech classification datasets demonstrates that the proposed framework is a promising methodology for analyzing and designing high-performing MCS for this task. MCS system obtained using the proposed framework significantly outperforms the combination of all models and the homogeneous and heterogeneous selection heuristics, demonstrating the importance of having a proper selection scheme. Source code, figures, and dataset splits can be found in the GitHub repository: https://github.com/Menelau/Hate-Speech-MCS.
Recent advances on prompt-tuning cast few-shot classification tasks as a masked language modeling problem. By wrapping input into a template and using a verbalizer which constructs a mapping between label space and label word space, prompt-tuning can achieve excellent results in zero-shot and few-shot scenarios. However, typical prompt-tuning needs a manually designed verbalizer which requires domain expertise and human efforts. And the insufficient label space may introduce considerable bias into the results. In this paper, we focus on eliciting knowledge from pretrained language models and propose a prototypical prompt verbalizer for prompt-tuning. Labels are represented by prototypical embeddings in the feature space rather than by discrete words. The distances between the embedding at the masked position of input and prototypical embeddings are used as classification criterion. For zero-shot settings, knowledge is elicited from pretrained language models by a manually designed template to form initial prototypical embeddings. For few-shot settings, models are tuned to learn meaningful and interpretable prototypical embeddings. Our method optimizes models by contrastive learning. Extensive experimental results on several many-class text classification datasets with low-resource settings demonstrate the effectiveness of our approach compared with other verbalizer construction methods. Our implementation is available at https://github.com/Ydongd/prototypical-prompt-verbalizer.
Multi-view learning accomplishes the task objectives of classification by leverag-ing the relationships between different views of the same object. Most existing methods usually focus on consistency and complementarity between multiple views. But not all of this information is useful for classification tasks. Instead, it is the specific discriminating information that plays an important role. Zhong Zhang et al. explore the discriminative and non-discriminative information exist-ing in common and view-specific parts among different views via joint non-negative matrix factorization. In this paper, we improve this algorithm on this ba-sis by using the cross entropy loss function to constrain the objective function better. At last, we implement better classification effect than original on the same data sets and show its superiority over many state-of-the-art algorithms.
Most adversarial attack defense methods rely on obfuscating gradients. These methods are successful in defending against gradient-based attacks; however, they are easily circumvented by attacks which either do not use the gradient or by attacks which approximate and use the corrected gradient. Defenses that do not obfuscate gradients such as adversarial training exist, but these approaches generally make assumptions about the attack such as its magnitude. We propose a classification model that does not obfuscate gradients and is robust by construction without assuming prior knowledge about the attack. Our method casts classification as an optimization problem where we "invert" a conditional generator trained on unperturbed, natural images to find the class that generates the closest sample to the query image. We hypothesize that a potential source of brittleness against adversarial attacks is the high-to-low-dimensional nature of feed-forward classifiers which allows an adversary to find small perturbations in the input space that lead to large changes in the output space. On the other hand, a generative model is typically a low-to-high-dimensional mapping. While the method is related to Defense-GAN, the use of a conditional generative model and inversion in our model instead of the feed-forward classifier is a critical difference. Unlike Defense-GAN, which was shown to generate obfuscated gradients that are easily circumvented, we show that our method does not obfuscate gradients. We demonstrate that our model is extremely robust against black-box attacks and has improved robustness against white-box attacks compared to naturally trained, feed-forward classifiers.
Applying standard machine learning approaches for classification can produce unequal results across different demographic groups. When then used in real-world settings, these inequities can have negative societal impacts. This has motivated the development of various approaches to fair classification with machine learning models in recent years. In this paper, we consider the problem of modifying the predictions of a blackbox machine learning classifier in order to achieve fairness in a multiclass setting. To accomplish this, we extend the 'post-processing' approach in Hardt et al. 2016, which focuses on fairness for binary classification, to the setting of fair multiclass classification. We explore when our approach produces both fair and accurate predictions through systematic synthetic experiments and also evaluate discrimination-fairness tradeoffs on several publicly available real-world application datasets. We find that overall, our approach produces minor drops in accuracy and enforces fairness when the number of individuals in the dataset is high relative to the number of classes and protected groups.
Existing methods for explaining black box learning models often focus on building local explanations of model behaviour for a particular data item. It is possible to create global explanations for all data items, but these explanations generally have low fidelity for complex black box models. We propose a new supervised manifold visualisation method, SLISEMAP, that simultaneously finds local explanations for all data items and builds a (typically) two-dimensional global visualisation of the black box model such that data items with similar local explanations are projected nearby. We provide a mathematical derivation of our problem and an open source implementation implemented using the GPU-optimised PyTorch library. We compare SLISEMAP to multiple popular dimensionality reduction methods and find that SLISEMAP is able to utilise labelled data to create embeddings with consistent local white box models. We also compare SLISEMAP to other model-agnostic local explanation methods and show that SLISEMAP provides comparable explanations and that the visualisations can give a broader understanding of black box regression and classification models.
The imbalanced classification problem turns out to be one of the important and challenging problems in data mining and machine learning. The performances of traditional classifiers will be severely affected by many data problems, such as class imbalanced problem, class overlap and noise. The Tomek-Link algorithm was only used to clean data when it was proposed. In recent years, there have been reports of combining Tomek-Link algorithm with sampling technique. The Tomek-Link sampling algorithm can effectively reduce the class overlap on data, remove the majority instances that are difficult to distinguish, and improve the algorithm classification accuracy. However, the Tomek-Links under-sampling algorithm only considers the boundary instances that are the nearest neighbors to each other globally and ignores the potential local overlapping instances. When the number of minority instances is small, the under-sampling effect is not satisfactory, and the performance improvement of the classification model is not obvious. Therefore, on the basis of Tomek-Link, a multi-granularity relabeled under-sampling algorithm (MGRU) is proposed. This algorithm fully considers the local information of the data set in the local granularity subspace, and detects the local potential overlapping instances in the data set. Then, the overlapped majority instances are eliminated according to the global relabeled index value, which effectively expands the detection range of Tomek-Links. The simulation results show that when we select the optimal global relabeled index value for under-sampling, the classification accuracy and generalization performance of the proposed under-sampling algorithm are significantly better than other baseline algorithms.
Classification problems are common in Computer Vision. Despite this, there is no dedicated work for the classification of beer bottles. As part of the challenge of the master course Deep Learning, a dataset of 5207 beer bottle images and brand labels was created. An image contains exactly one beer bottle. In this paper we present a deep learning model which classifies pictures of beer bottles in a two step approach. As the first step, a Faster-R-CNN detects image sections relevant for classification independently of the brand. In the second step, the relevant image sections are classified by a ResNet-18. The image section with the highest confidence is returned as class label. We propose a model, with which we surpass the classic one step transfer learning approach and reached an accuracy of 99.86% during the challenge on the final test dataset. We were able to achieve 100% accuracy after the challenge ended
Cardiovascular disease is a large worldwide healthcare issue; symptoms often present suddenly with minimal warning. The electrocardiogram (ECG) is a fast, simple and reliable method of evaluating the health of the heart, by measuring electrical activity recorded through electrodes placed on the skin. ECGs often need to be analyzed by a cardiologist, taking time which could be spent on improving patient care and outcomes. Because of this, automatic ECG classification systems using machine learning have been proposed, which can learn complex interactions between ECG features and use this to detect abnormalities. However, algorithms built for this purpose often fail to generalize well to unseen data, reporting initially impressive results which drop dramatically when applied to new environments. Additionally, machine learning algorithms suffer a "black-box" issue, in which it is difficult to determine how a decision has been made. This is vital for applications in healthcare, as clinicians need to be able to verify the process of evaluation in order to trust the algorithm. This paper proposes a method for visualizing model decisions across each class in the MIT-BIH arrhythmia dataset, using adapted saliency maps averaged across complete classes to determine what patterns are being learned. We do this by building two algorithms based on state-of-the-art models. This paper highlights how these maps can be used to find problems in the model which could be affecting generalizability and model performance. Comparing saliency maps across complete classes gives an overall impression of confounding variables or other biases in the model, unlike what would be highlighted when comparing saliency maps on an ECG-by-ECG basis.
Currently, publicly available models for website classification do not offer an embedding method and have limited support for languages beyond English. We release a dataset of more than two million category-labeled websites in 92 languages collected from Curlie, the largest multilingual human-edited Web directory. The dataset contains 14 website categories aligned across languages. Alongside it, we introduce Homepage2Vec, a machine-learned pre-trained model for classifying and embedding websites based on their homepage in a language-agnostic way. Homepage2Vec, thanks to its feature set (textual content, metadata tags, and visual attributes) and recent progress in natural language representation, is language-independent by design and generates embedding-based representations. We show that Homepage2Vec correctly classifies websites with a macro-averaged F1-score of 0.90, with stable performance across low- as well as high-resource languages. Feature analysis shows that a small subset of efficiently computable features suffices to achieve high performance even with limited computational resources. We make publicly available the curated Curlie dataset aligned across languages, the pre-trained Homepage2Vec model, and libraries
A lot of prognostication methodologies have been formulated for early detection of Polycystic Ovary Syndrome also known as PCOS using Machine Learning. PCOS is a binary classification problem. Dimensionality Reduction methods impact the performance of Machine Learning to a greater extent and using a Supervised Dimensionality Reduction method can give us a new edge to tackle this problem. In this paper we present Discriminant Analysis in different dimensions with Linear and Quadratic form for binary classification along with metrics. We were able to achieve good accuracy and less variation with Discriminant Analysis as compared to many commonly used classification algorithms with training accuracy reaching 97.37% and testing accuracy of 95.92% using Quadratic Discriminant Analysis. Paper also gives the analysis of data with visualizations for deeper understanding of problem.
Linear and Quadratic Discriminant Analysis are well-known classical methods but can heavily suffer from non-Gaussian distributions and/or contaminated datasets, mainly because of the underlying Gaussian assumption that is not robust. To fill this gap, this paper presents a new robust discriminant analysis where each data point is drawn by its own arbitrary Elliptically Symmetrical (ES) distribution and its own arbitrary scale parameter. Such a model allows for possibly very heterogeneous, independent but non-identically distributed samples. After deriving a new decision rule, it is shown that maximum-likelihood parameter estimation and classification are very simple, fast and robust compared to state-of-the-art methods.
In the tasks of image aesthetic quality evaluation, it is difficult to reach both the high score area and low score area due to the normal distribution of aesthetic datasets. To reduce the error in labeling and solve the problem of normal data distribution, we propose a new aesthetic mixed dataset with classification and regression called AMD-CR, and we train a meta reweighting network to reweight the loss of training data differently. In addition, we provide a training strategy acccording to different stages, based on pseudo labels of the binary classification task, and then we use it for aesthetic training acccording to different stages in classification and regression tasks. In the construction of the network structure, we construct an aesthetic adaptive block (AAB) structure that can adapt to any size of the input images. Besides, we also use the efficient channel attention (ECA) to strengthen the feature extracting ability of each task. The experimental result shows that our method improves 0.1112 compared with the conventional methods in SROCC. The method can also help to find best aesthetic path planning for unmanned aerial vehicles (UAV) and vehicles.
In this paper we present the Amharic Speech Emotion Dataset (ASED), which covers four dialects (Gojjam, Wollo, Shewa and Gonder) and five different emotions (neutral, fearful, happy, sad and angry). We believe it is the first Speech Emotion Recognition (SER) dataset for the Amharic language. 65 volunteer participants, all native speakers, recorded 2,474 sound samples, two to four seconds in length. Eight judges assigned emotions to the samples with high agreement level (Fleiss kappa = 0.8). The resulting dataset is freely available for download. Next, we developed a four-layer variant of the well-known VGG model which we call VGGb. Three experiments were then carried out using VGGb for SER, using ASED. First, we investigated whether Mel-spectrogram features or Mel-frequency Cepstral coefficient (MFCC) features work best for Amharic. This was done by training two VGGb SER models on ASED, one using Mel-spectrograms and the other using MFCC. Four forms of training were tried, standard cross-validation, and three variants based on sentences, dialects and speaker groups. Thus, a sentence used for training would not be used for testing, and the same for a dialect and speaker group. The conclusion was that MFCC features are superior under all four training schemes. MFCC was therefore adopted for Experiment 2, where VGGb and three other existing models were compared on ASED: RESNet50, Alex-Net and LSTM. VGGb was found to have very good accuracy (90.73%) as well as the fastest training time. In Experiment 3, the performance of VGGb was compared when trained on two existing SER datasets, RAVDESS (English) and EMO-DB (German) as well as on ASED (Amharic). Results are comparable across these languages, with ASED being the highest. This suggests that VGGb can be successfully applied to other languages. We hope that ASED will encourage researchers to experiment with other models for Amharic SER.
The assessment of craniofacial deformities requires patient data which is sparsely available. Statistical shape models provide realistic and synthetic data enabling comparisons of existing methods on a common dataset. We build the first publicly available statistical 3D head model of craniosynostosis patients and the first model focusing on infants younger than 1.5 years. We further present a shape-model-based classification pipeline to distinguish between three different classes of craniosynostosis and a control group on photogrammetric surface scans. To the best of our knowledge, our study uses the largest dataset of craniosynostosis patients in a classification study for craniosynostosis and statistical shape modeling to date. We demonstrate that our shape model performs similar to other statistical shape models of the human head. Craniosynostosis-specific pathologies are represented in the first eigenmodes of the model. Regarding the automatic classification of craniosynostis, our classification approach yields an accuracy of 97.8%, comparable to other state-of-the-art methods using both computed tomography scans and stereophotogrammetry. Our publicly available, craniosynostosis-specific statistical shape model enables the assessment of craniosynostosis on realistic and synthetic data. We further present a state-of-the-art shape-model-based classification approach for a radiation-free diagnosis of craniosynostosis.
It is observed that high classification performance is achieved for one- and two-dimensional signals by using deep learning methods. In this context, most researchers have tried to classify hyperspectral images by using deep learning methods and classification success over 90% has been achieved for these images. Deep neural networks (DNN) actually consist of two parts: i) Convolutional neural network (CNN) and ii) fully connected neural network (FCNN). While CNN determines the features, FCNN is used in classification. In classification of the hyperspectral images, it is observed that almost all of the researchers used 2D or 3D convolution filters on the spatial data beside spectral data (features). It is convenient to use convolution filters on images or time signals. In hyperspectral images, each pixel is represented by a signature vector which consists of individual features that are independent of each other. Since the order of the features in the vector can be changed, it doesn't make sense to use convolution filters on these features as on time signals. At the same time, since the hyperspectral images do not have a textural structure, there is no need to use spatial data besides spectral data. In this study, hyperspectral images of Indian pines, Salinas, Pavia centre, Pavia university and Botswana are classified by using only fully connected neural network and the spectral data with one dimensional. An average accuracy of 97.5% is achieved for the test sets of all hyperspectral images.
The "Roaring 20s" of visual recognition began with the introduction of Vision Transformers (ViTs), which quickly superseded ConvNets as the state-of-the-art image classification model. A vanilla ViT, on the other hand, faces difficulties when applied to general computer vision tasks such as object detection and semantic segmentation. It is the hierarchical Transformers (e.g., Swin Transformers) that reintroduced several ConvNet priors, making Transformers practically viable as a generic vision backbone and demonstrating remarkable performance on a wide variety of vision tasks. However, the effectiveness of such hybrid approaches is still largely credited to the intrinsic superiority of Transformers, rather than the inherent inductive biases of convolutions. In this work, we reexamine the design spaces and test the limits of what a pure ConvNet can achieve. We gradually "modernize" a standard ResNet toward the design of a vision Transformer, and discover several key components that contribute to the performance difference along the way. The outcome of this exploration is a family of pure ConvNet models dubbed ConvNeXt. Constructed entirely from standard ConvNet modules, ConvNeXts compete favorably with Transformers in terms of accuracy and scalability, achieving 87.8% ImageNet top-1 accuracy and outperforming Swin Transformers on COCO detection and ADE20K segmentation, while maintaining the simplicity and efficiency of standard ConvNets.
For several decades, planet identification has only been performed by astronomical experts and researchers with the help of specialized equipment. With the advent of computational methods and access to satellite data from space missions, this trend has changed. For instance, NASA’s Exoplanet Exploration program has provided us with vast amounts of data on celestial objects to assist in space exploration. One such mission of interest is the Kepler mission. Over 4000 such transiting exoplanets have been identified since the mission commenced in 2007. It has provided us with an extensive database of discoveries that help in computing planet occurrence rates as a function of an object’s parameters such as the size, insolation flux, star type and orbital period. This information is catalogued in the Cumulative Kepler Object of Information dataset. Four basic models have been compared. Namely, Support Vector Machines, Random Forest Classifiers, AdaBoost and Deep Neural Networks. The AdaBoost classifier was selected as the optimum machine learning model and returned an F-1 score of 0.98.