Locally Alignment Driven Multi Kernel SVM for Robust Prediction on Heterogeneous Biomedical Data

Objective. Learning to classify cardiac abnormalities requires large and high-quality labeled datasets, which is a challenge in medical applications. Small datasets from various sources are often aggregated to meet this requirement, resulting in a final dataset prone to label noise due to inter- and intra-observer variability and different expertise. It is well known that label noise can affect the performance and generalizability of the trained models. In this work, we explore the impact of label noise and self-learning label correction on the classification of cardiac abnormalities on large heterogeneous datasets of electrocardiogram (ECG) signals. Approach. A state-of-the-art self-learning multi-class label correction method for image classification is adapted to learn a multi-label classifier for electrocardiogram signals. We evaluated our performance using 5-fold cross-validation on the publicly available PhysioNet/Computing in Cardiology (CinC) 2021 Challenge data, with full and reduced sets of leads. Due to the unknown label noise in the testing set, we tested our approach on the MNIST dataset. We investigated the performance under different levels of structured label noise for both datasets. Main results. Under high levels of noise, the cross-validation results of self-learning label correction show an improvement of approximately 3% in the challenge score for the PhysioNet/CinC 2021 Challenge dataset and an improvement in accuracy of 5% and reduction of the expected calibration error of 0.03 for the MNIST dataset. We demonstrate that self-learning label correction can be used to effectively deal with the presence of unknown label noise, also when using a reduced number of ECG leads.

Black-box neural networks are inherently inscrutable, and their widespread use has triggered significant societal issues in crucial areas such as healthcare, finance and safety. In these high-stakes decision-making domains, the deployment of machine learning algorithms requires not only prediction accuracy but also their interpretability and robustness against data distribution shifts, such as outliers, label noise, and category imbalance. In this work, we propose a novel Meta-weighted Sparse Neural Additive Model (MSpNAM), which offers robustness through an efficient bilevel weighting policy and inherits strong explainability and representation capabilities from the additive modeling strategy. Furthermore, empirical results across multiple synthetic and real datasets, under various distribution shifts, demonstrate that MSpNAM can scale effectively and achieve superior performance in terms of robustness, interpretability, and anti-forgetting compared to some of the latest baselines.

LiDAR-based global localization provides accurate robot pose estimates against a prior map. Existing deep-learning methods, however, demand heavy computation and long training or inference times and degrade sharply when faced with domain shifts. This letter presents LighterBEV, a lightweight, fast, and generalizable localization method. An Informative Compression Module achieves a fourfold reduction in local-feature dimensionality while improving accuracy. We further integrate online learning to enable rapid post-deployment adaptation, mitigating degradation under distribution shift. Extensive experiments on four large-scale datasets show that LighterBEV achieves state-of-the-art performance with limited training data, maintains high accuracy under domain shift, and runs in real time on resource-constrained hardware-supporting both inference and online updates. To our knowledge, LighterBEV is the first LiDAR global localization approach to incorporate online learning for automatic adaptation to new environments, thereby narrowing the domain gap.

A key challenge with supervised learning (e.g., image classification) is the shift of data distribution and domain from training to testing datasets, so-called “domain shift” (or “distribution shift”), which usually leads to a reduction of model accuracy. Various meta-learning approaches have been proposed to prevent the accuracy loss by learning an adaptable model with training data, and adapting it to test time data from a new data domain. However, when the domain shift occurs in multiple domain dimensions (e.g., images may be transformed by rotations, transitions, and expansions), the average predictive power of the adapted model will deteriorate. To tackle this problem, we propose a domain disentangled meta-learning (DDML) framework. DDML disentangles the data domain by dimensions, learns the representations of domain dimensions independently, and adapts to the domain of test time data. We evaluate our DDML on image classification problems using three datasets with distribution shifts over multiple domain dimensions. Comparing to various baselines in meta-learning and empirical risk minimization, our DDML approach achieves consistently higher classification accuracy with the test time data. These results demonstrate that domain disentanglement reduces the complexity of the model adaptation, thus increases the model generalizability, and prevents it from overfitting.

Wearable technologies enable continuous monitoring of various health metrics, such as physical activity, heart rate, sleep, and stress levels. A key challenge with wearable data is obtaining quality labels. Unlike modalities like video where the videos themselves can be effectively used to label objects or events, wearable data do not contain obvious cues about the physical manifestation of the users and usually require rich metadata. As a result, label noise can become an increasingly thorny issue when labeling wearable data. In this paper, we propose a novel solution to address noisy label learning, entitled Few-Shot Human-in-the-Loop Refinement (FHLR). Our method initially learns a seed model using weak labels. Next, it fine-tunes the seed model using a handful of expert corrections. Finally, it achieves better generalizability and robustness by merging the seed and fine-tuned models via weighted parameter averaging. We evaluate our approach on four challenging tasks and datasets, and compare it against eight competitive baselines designed to deal with noisy labels. We show that FHLR achieves significantly better performance when learning from noisy labels and achieves state-of-the-art by a large margin, with up to accuracy improvement under symmetric and asymmetric noise. Notably, we find that FHLR is particularly robust to increased label noise, unlike prior works that suffer from severe performance degradation. Our work not only achieves better generalization in high-stakes health sensing benchmarks but also sheds light on how noise affects commonly-used models.

Medical image segmentation tasks hitherto have achieved excellent progresses with large-scale datasets, which empowers us to train potent deep convolutional neural networks (DCNNs). However, labeling such large-scale datasets is laborious and error-prone, which leads the noisy (or incorrect) labels to be an ubiquitous problem in the real-world scenarios. In addition, data collected from different sites usually exhibit significant data distribution shift (or domain shift). As a result, noisy label and domain shift become two common problems in medical imaging application scenarios, especially in medical image segmentation, which degrade the performance of deep learning models significantly. In this paper, we identify a novel problem hidden in medical image segmentation, which is unsupervised domain adaptation on noisy labeled data, and propose a novel algorithm named "Self-Cleansing Unsupervised Domain Adaptation" (S-CDUA) to address such issue. S-CUDA sets up a realistic scenario to solve the above problems simultaneously where training data (i.e., source domain) not only shows domain shift w.r.t. unsupervised test data (i.e., target domain) but also contains noisy labels. The key idea of S-CUDA is to learn noise-excluding and domain invariant knowledge from noisy supervised data, which will be applied on the highly corrupted data for label cleansing and further data-recycling, as well as on the test data with domain shift for supervised propagation. To this end, we propose a novel framework leveraging noisy-label learning and domain adaptation techniques to cleanse the noisy labels and learn from trustable clean samples, thus enabling robust adaptation and prediction on the target domain. Specifically, we train two peer adversarial networks to identify high-confidence clean data and exchange them in companions to eliminate the error accumulation problem and narrow the domain gap simultaneously. In the meantime, the high-confidence noisy data are detected and cleansed in order to reuse the contaminated training data. Therefore, our proposed method can not only cleanse the noisy labels in the training set but also take full advantage of the existing noisy data to update the parameters of the network. For evaluation, we conduct experiments on two popular datasets (REFUGE and Drishti-GS) for optic disc (OD) and optic cup (OC) segmentation, and on another public multi-vendor dataset for spinal cord gray matter (SCGM) segmentation. Experimental results show that our proposed method can cleanse noisy labels efficiently and obtain a model with better generalization performance at the same time, which outperforms previous state-of-the-art methods by large margin. Our code can be found at https://github.com/zzdxjtu/S-cuda.

We address an important and practical problem facing deep-learning-based image classification: class incremental learning with a large domain shift. Most previous efforts on class incremental learning focus on one aspect of the problem, i.e., learning to classify additional new classes (with a little shift). However, in the real world, when new classes are added, the domain changes simultaneously (with a large domain shift). To obtain a model that is robust to these situations, we need to consider incrementally learning not only new labels but also domain-shifted labels. We target a continual and simultaneous shift of class and domain distribution and propose a new incremental learning method named Momentum Contrastive learning enhancing Orthogonality of Negative pairs (MoCo-ON). We employ a momentum encoder framework augmented with rehearsal memory to mitigate the risk of forgetting while leveraging contrastive learning to extract versatile features capable of adapting to the progressively shifting domain. Specifically, when training with a knowledge distillation loss, we introduce a novel supervised contrastive loss designed to closely embed positive pairs of the same class, even in the presence of a substantial domain gap. Additionally, we leverage feature embedding from momentum encoders for exemplar selection, aiming to mitigate the risk of forgetting previously acquired information from earlier tasks. We conduct comprehensive experiments involving inter-domain shifted class incremental learning scenarios using widely adopted datasets commonly employed for studying domain generalization in image classification. Our proposed method consistently outperforms other methods by a significant margin.

Recent neural architectures in named entity recognition (NER) have yielded state-of-the-art performance on single domain data such as newswires. However, they still suffer from (i) requiring massive amounts of training data to avoid overfitting; (ii) huge performance degradation when there is a domain shift in the data distribution between training and testing. In this paper, we investigate the problem of domain adaptation for NER under homogeneous and heterogeneous settings. We propose MetaNER, a novel meta-learning approach for domain adaptation in NER. Specifically, MetaNER incorporates meta-learning and adversarial training strategies to encourage robust, general and transferable representations for sequence labeling. The key advantage of MetaNER is that it is capable of adapting to new unseen domains with a small amount of annotated data from those domains. We extensively evaluate MetaNER on multiple datasets under homogeneous and heterogeneous settings. The experimental results show that MetaNER achieves state-of-the-art performance against eight baselines. Impressively, MetaNER surpasses the in-domain performance using only 16.17% and 34.76% of target domain data on average for homogeneous and heterogeneous settings, respectively.

We release a new dataset containing invasive coronary angiograms for the coronary dominance classification task, an essential aspect of SYNTAX score estimation for assessing the severity of coronary artery disease. The dataset contains 1,574 angiographic studies, a set of X-ray multi-view videos classified as right or left dominant. The dataset has three parts: main, real distribution, and domain shift. Each study in the main dataset falls into five categories: normal, poor quality, artifact, high uncertainty, or right coronary artery (RCA) occlusion. These tags help classify the dominance more accurately and allow for uncertainty estimation and outlier detection. In the real distribution part, there are 400 randomly selected studies. In contrast the domain shift part includes studies from an interventional angiography system that differs from that used in the main and real distribution parts.

Among the other cancer types, the brain tumor is one the leading cause of cancer across globe. If the tumor is properly identified at an earlier stage, then the chances of the survival can be increased. To categorize the brain tumor there are several factors including texture, type and location of brain tumor. We proposed a novel reconstruction independent component analysis (RICA) feature extraction method to detect multi-class brain tumor types (pituitary, meningioma, and glioma). We then employed the robust machine learning techniques as support vector machine (SVM) with quadratic and linear kernels and linear discriminant analysis (LDA). For training and testing of the data validation, a 10-fold cross validation was employed. For the multi-class classification, the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), accuracy and AUC were, respectively, 97.78%, 100%, 100%, 99.07, 99.34% and 0.9892 to detect pituitary using SVM Cubic followed by meningioma with accuracy (96.96%0, AUC (0.9348) and glioma with accuracy (95.88%), AUC (0.9635). The findings indicates that RICA feature based proposed methodology has more potential to detect the multiclass brain tumor types for improving diagnostic efficiency and can further improve the prediction accuracy to achieve the clinical outcomes.

This study aims to predict a student’s graduation on time (GOT) using machine learning algorithms. We applied five different machine learning algorithms, namely Random Forest, Support Vector Machine (Linear Kernel), Support Vector Machine (Polynomial Kernel), K-Nearest Neighbors, and Naïve Bayes. These algorithms were tested using 10-fold cross validation and simulated various parameter tuning values. The results show that the Random Forest algorithm produces the best accuracy and kappa statistics values, so this algorithm is suitable for modeling predictive data of students graduating on time. This predictive model is expected to be useful for higher education management in designing their strategies to assist and improve student academic performance weaknesses in order to achieve graduation on time.

Dietary assessment has proven to be effective to evaluate the dietary intake of patients with diabetes and obesity. The traditional approach of accessing the dietary intake is to conduct a 24-hour dietary recall, a structured interview designed to obtain information on food categories and volume consumed by the participants. Due to unconscious biases in this kind of self-reporting approaches, many research studies have explored the use of vision-based approaches to provide accurate and objective assessments. Despite the promising results of food recognition by deep neural networks, there still exist several hurdles in deep learning-based food volume estimation ranging from domain shift between synthetic and raw 3D models, shape completion ambiguity and lack of large-scale paired training dataset. Therefore, this paper proposed an intelligent nutritional assessment approach via weakly-supervised point cloud completion, which aims to close the reality gap in 3D point cloud completion tasks and address the targeted challenges. Then the volume can be easily estimated from the completed representation of the food. Another major merit of our system is that it can be used to estimate the volume of handheld food items without requiring the constraints including placing the food items on a table or next to fiducial markers, which facilitates the implementation on both wearable and handheld cameras. Comprehensive experiments have been carried out on major benchmark datasets and self-constructed volume-annotated dataset respectively, in which the proposed method demonstrates comparable results with several strong fully-supervised baseline methods and shows superior completion ability in handling food volume estimation.

The visual similarity among Calathea varieties—especially in leaf color, pattern, and texture, often causes misidentification by the general public and even horticulturists. This research proposes an automated classification system for five Calathea varieties (Black Lipstick, Corona, Crimson, Medallion, and Pink Jessy) using Support Vector Machine (SVM) based on combined HSV (Hue, Saturation, Value) and HOG (Histogram of Oriented Gradients) features. A dataset of 1,000 leaf images was collected under controlled conditions (white background, fixed distance, consistent lighting) and processed through preprocessing, HSV-based color segmentation, grayscale conversion, and HOG feature extraction. The SVM model with a linear kernel was evaluated using 5-fold cross-validation. The system achieved an overall average accuracy of 93.52%, with Calathea Corona showing the highest accuracy (96.50%) and precision (95.12%), while Calathea Black Lipstick recorded the lowest precision (67.50%) but the highest recall (92.00%). These results demonstrate that the fusion of color and shape features enhances classification performance, although complex leaf patterns can still pose challenges in precision. This approach offers a practical and objective solution for Calathea identification in horticultural applications.

The aim of this study was to evaluate various radiomics-based machine learning classification models using the apparent diffusion coefficient (ADC) and cerebral blood flow (CBF) maps for differentiating between low-grade gliomas (LGGs) and high-grade gliomas (HGGs). Fifty-two glioma patients, including 18 LGGs (grade II) and 34 HGGs (grade III/IV), were examined using a 3.0-T magnetic resonance scanner. The ADC and CBF maps were obtained from diffusion-weighted imaging and pseudo-continuous arterial spin labeling perfusion-weighted imaging, respectively. A total of 91 radiomic features were extracted from each of the tumor volume on the ADC and CBF maps. We constructed 4 types of machine learning classifiers based on (1) least absolute shrinkage and selection operator regularized logistic regression (LASSO-LR), (2) random forest (RF), (3) support vector machine (SVM) with the radial basis function kernel (SVM-RBF), and (4) SVM with the linear kernel (SVM-L). A training set with 36 gliomas (70%) was used to select the important radiomic features and train each model using 5-fold cross-validation. The remaining 16 gliomas (30%) were used as a test set. Receiver operating characteristic analysis was performed to evaluate the model performance. A radiomic feature, ADC first-order-based skewness, was selected as an important variable in all classification models. According to the receiver operating characteristic analysis, the areas under the curve of the LASSO-LR, RF, SVM-RBF, and SVM-L models for the training set were 0.965, 1.000, 0.979, and 0.969, respectively. For the test set, the areas under the curve of the LASSO-LR, RF, SVM-RBF, and SVM-L models were 0.883, 0.917, 0.717, and 0.917, respectively. All classification models showed sufficient diagnostic performance on the test set. Radiomics-based machine learning classifiers using the quantitative ADC and CBF maps are useful for differentiating HGGs from LGGs.

In this paper, an automatic brain tumor diagnosis system is presented using a new threshold-based segmentation method. The proposed segmentation method is based on collaboration of beta mixture model and learning automata (LA). This segmentation method approximates histogram of a given MR image through mixture of Beta functions whose optimized parameters are determined through LA tool. Each Beta function shows one pixel class, and the threshold values which used for segmentation are obtained via intersection of two adjacent Beta functions. Feature extraction is based on statistical features. Moreover, support vector machine (SVM), K-nearest neighbor (KNN) and decision tree (DT) are applied as binary classifiers. In order to evaluate the performance, the proposed method is analyzed on set of 79 MR images from TCIA dataset and Harvard Medical School. The results of the experiments show that the proposed segmentation method presents superior average values for image similarity indices Dice and Jaccard. Moreover, the best accuracy of the presented brain tumor diagnosis system is obtained using SVM classifier with linear kernel, which is more than 98% in 10-fold cross-validation.