Supervised Chromatin Loop Detection Using Peakachu Version 2.

Peakachu is a supervised-learning-based approach that identifies chromatin loops from chromatin contact data. Here, we present Peakachu version 2, an updated version that significantly improves extensibility, usability, and computational efficiency compared to its predecessor. It features pretrained models tailored for a wide range of experimental platforms, such as Hi-C, Micro-C, ChIA-PET, HiChIP, HiCAR, and TrAC-loop. This chapter offers a step-by-step tutorial guiding users through the process of training Peakachu models from scratch and utilizing pretrained models to predict chromatin loops across various platforms.

Applying support vector machines to a diagnostic classification model for polytomous attributes in small-sample contexts.

Over several years, the evaluation of polytomous attributes in small-sample settings has posed a challenge to the application of cognitive diagnosis models. To enhance classification precision, the support vector machine (SVM) was introduced for estimating polytomous attribution, given its proven feasibility for dichotomous cases. Two simulation studies and an empirical study assessed the impact of various factors on SVM classification performance, including training sample size, attribute structures, guessing/slipping levels, number of attributes, number of attribute levels, and number of items. The results indicated that SVM outperformed the pG-DINA model in classification accuracy under dependent attribute structures and small sample sizes. SVM performance improved with an increased number of items but declined with higher guessing/slipping levels, more attributes, and more attribute levels. Empirical data further validated the application and advantages of SVMs.

A self-supervised deep Riemannian representation to classify parkinsonian fixational patterns.

Parkinson's disease (PD) is the second most prevalent neurodegenerative disorder, and it remains incurable. Currently there is no definitive biomarker for detecting PD, measuring its severity, or monitoring of treatments. Recently, oculomotor fixation abnormalities have emerged as a sensitive biomarker to discriminate Parkinsonian patterns from a control population, even at early stages. For oculomotor analysis, current experimental setups use invasive and restrictive capture protocols that limit the transfer in clinical routine. Alternatively, computational approaches to support the PD diagnosis are strictly based on supervised strategies, depending of large labeled data, and introducing an inherent expert-bias. This work proposes a self-supervised architecture based on Riemannian deep representation to learn oculomotor fixation patterns from compact descriptors. Firstly, deep convolutional features are recovered from oculomotor fixation video slices, and then encoded in compact symmetric positive matrices (SPD) to summarize second-order relationships. Each SPD input matrix is projected onto a Riemannian encoder until obtain a SPD embedding. Then, a Riemannian decoder reconstructs SPD matrices while preserving the geometrical manifold structure. The proposed architecture successfully recovers geometric patterns in the embeddings without any label diagnosis supervision, and demonstrates the capability to be discriminative regarding PD patterns. In a retrospective study involving 13 healthy adults and 13 patients diagnosed with PD, the proposed Riemannian representation achieved an average accuracy of 95.6% and an AUC of 99% during a binary classification task using a Support Vector Machine.

Efficient multi-view fusion and flexible adaptation to view missing in cardiovascular system signals.

The progression of deep learning and the widespread adoption of sensors have facilitated automatic multi-view fusion (MVF) about the cardiovascular system (CVS) signals. However, prevalent MVF model architecture often amalgamates CVS signals from the same temporal step but different views into a unified representation, disregarding the asynchronous nature of cardiovascular events and the inherent heterogeneity across views, leading to catastrophic view confusion. Efficient training strategies specifically tailored for MVF models to attain comprehensive representations need simultaneous consideration. Crucially, real-world data frequently arrives with incomplete views, an aspect rarely noticed by researchers. Thus, the View-Centric Transformer (VCT) and Multitask Masked Autoencoder (M2AE) are specifically designed to emphasize the centrality of each view and harness unlabeled data to achieve superior fused representations. Additionally, we systematically define the missing-view problem for the first time and introduce prompt techniques to aid pretrained MVF models in flexibly adapting to various missing-view scenarios. Rigorous experiments involving atrial fibrillation detection, blood pressure estimation, and sleep staging-typical health monitoring tasks-demonstrate the remarkable advantage of our method in MVF compared to prevailing methodologies. Notably, the prompt technique requires finetuning <3 % of the entire model's data, substantially fortifying the model's resilience to view missing while circumventing the need for complete retraining. The results demonstrate the effectiveness of our approaches, highlighting their potential for practical applications in cardiovascular health monitoring. Codes and models are released at URL.

Demystifying unsupervised learning: how it helps and hurts.

Humans and machines rarely have access to explicit external feedback or supervision, yet manage to learn. Most modern machine learning systems succeed because they benefit from unsupervised data. Humans are also expected to benefit and yet, mysteriously, empirical results are mixed. Does unsupervised learning help humans or not? Here, we argue that the mixed results are not conflicting answers to this question, but reflect that humans self-reinforce their predictions in the absence of supervision, which can help or hurt depending on whether predictions and task align. We use this framework to synthesize empirical results across various domains to clarify when unsupervised learning will help or hurt. This provides new insights into the fundamentals of learning with implications for instruction and lifelong learning.

Ptychographic phase retrieval via a deep-learning-assisted iterative algorithm.

Ptychography is a powerful computational imaging technique with microscopic imaging capability and adaptability to various specimens. To obtain an imaging result, it requires a phase-retrieval algorithm whose performance directly determines the imaging quality. Recently, deep neural network (DNN)-based phase retrieval has been proposed to improve the imaging quality from the ordinary model-based iterative algorithms. However, the DNN-based methods have some limitations because of the sensitivity to changes in experimental conditions and the difficulty of collecting enough measured specimen images for training the DNN. To overcome these limitations, a ptychographic phase-retrieval algorithm that combines model-based and DNN-based approaches is proposed. This method exploits a DNN-based denoiser to assist an iterative algorithm like ePIE in finding better reconstruction images. This combination of DNN and iterative algorithms allows the measurement model to be explicitly incorporated into the DNN-based approach, improving its robustness to changes in experimental conditions. Furthermore, to circumvent the difficulty of collecting the training data, it is proposed that the DNN-based denoiser be trained without using actual measured specimen images but using a formula-driven supervised approach that systemically generates synthetic images. In experiments using simulation based on a hard X-ray ptychographic measurement system, the imaging capability of the proposed method was evaluated by comparing it with ePIE and rPIE. These results demonstrated that the proposed method was able to reconstruct higher-spatial-resolution images with half the number of iterations required by ePIE and rPIE, even for data with low illumination intensity. Also, the proposed method was shown to be robust to its hyperparameters. In addition, the proposed method was applied to ptychographic datasets of a Simens star chart and ink toner particles measured at SPring-8 BL24XU, which confirmed that it can successfully reconstruct images from measurement scans with a lower overlap ratio of the illumination regions than is required by ePIE and rPIE.

Bootstrap each lead's latent: A novel method for self-supervised learning of multilead electrocardiograms.

BACKGROUND AND OBJECTIVE: Electrocardiogram (ECG) is one of the most important diagnostic tools for cardiovascular diseases (CVDs). Recent studies show that deep learning models can be trained using labeled ECGs to achieve automatic detection of CVDs, assisting cardiologists in diagnosis. However, the deep learning models heavily rely on labels in training, while manual labeling is costly and time-consuming. This paper proposes a new self-supervised learning (SSL) method for multilead ECGs: bootstrap each lead's latent (BELL) to reduce the reliance and boost model performance in various tasks, especially when training data are insufficient. METHOD: BELL is a variant of the well-known bootstrap your own latent (BYOL). The BELL aims to learn prior knowledge from unlabeled ECGs by pretraining, benefitting downstream tasks. It leverages the characteristics of multilead ECGs. First, BELL uses the multiple-branch skeleton, which is more effective in processing multilead ECGs. Moreover, it proposes intra-lead and inter-lead mean square error (MSE) to guide pretraining, and their fusion can result in better performances. Additionally, BELL inherits the main advantage of the BYOL: No negative pair is used in pretraining, making it more efficient. RESULTS: In most cases, BELL surpasses previous works in the experiments. More importantly, the pretraining improves model performances by 0.69% ∼ 8.89% in downstream tasks when only 10% of training data are available. Furthermore, BELL shows excellent adaptability to uncurated ECG data from a real-world hospital. Only slight performance degradation occurs (<1% in most cases) when using these data. CONCLUSION: The results suggest that the BELL can alleviate the reliance on manual ECG labels from cardiologists, a critical bottleneck of the current deep learning-based models. In this way, the BELL can also help deep learning extend its application on automatic ECG analysis, reducing the cardiologists' burden in real-world diagnosis.

G2LCPS: End-to-end semi-supervised landmark prediction with global-to-local cross pseudo supervision for airway difficulty assessment.

Difficult tracheal intubation is a major cause of anesthesia-related injuries, including brain damage and death. While deep neural networks have improved difficult airways (DA) predictions over traditional assessment methods, existing models are often black boxes, making them difficult to trust in critical medical settings. Traditional DA assessment relies on facial and neck features, but detecting neck landmarks is particularly challenging. This paper introduces a novel semi-supervised method for landmark prediction, namely G2LCPS, which leverages hierarchical filters and cross-supervised signals. The novelty lies in ensuring that the networks select good unlabeled samples at the image level and generate high-quality pseudo heatmaps at the pixel level for cross-pseudo supervision. The extended versions of the public AFLW, CFP, CPLFW and CASIA-3D FaceV1 face datasets and show that G2LCPS achieves superior performance compared to other state-of-the-art semi-supervised methods, achieving the lowest normalized mean error (NME) of 3.588 when only 1/8 of data is labeled. Notably, the inclusion of the local filter improved the prediction by at least 0.199 NME, whereas the global filter contributed an additional improvement of at least 0.216 NME. These findings underscore the effectiveness of our approach, particularly in scenarios with limited labeled data, and suggest that G2LCPS can significantly enhance the reliability and accuracy of DA predictions in clinical practice. The results highlight the potential of our method to improve patient safety by providing more trustworthy and precise predictions for difficult airway management.

Enhancing dental caries classification in CBCT images by using image processing and self-supervised learning.

Diagnosing dental caries poses a significant challenge in dentistry, necessitating precise and early detection for effective management. This study utilizes Self-Supervised Learning (SSL) tasks to improve the classification of dental caries in Cone Beam Computed Tomography (CBCT) images, employing the International Caries Detection and Assessment System (ICDAS). Faced with the challenge of scarce annotated medical images, our research employs SSL to utilize unlabeled data, thereby improving model performance. We have developed a pipeline incorporating unlabeled data extraction from CBCT exams and subsequent model training using SSL tasks. A distinctive aspect of our approach is the integration of image processing techniques with SSL tasks, along with exploring the necessity for unlabeled data. Our research aims to identify the most effective image processing techniques for data extraction, the most efficient deep learning architectures for caries classification, the impact of unlabeled dataset sizes on model performance, and the comparative effectiveness of different SSL approaches in this domain. Among the tested architectures, ResNet-18, combined with the SimCLR task, demonstrated an average F1-score macro of 88.42%, Precision macro of 90.44%, and Sensitivity macro of 86.67%, reaching a 5.5% increase in F1-score compared to models using only deep learning architecture. These results suggest that SSL can significantly enhance the accuracy and efficiency of caries classification in CBCT images.

Self-adaptive label discovery and multi-view fusion for complementary label learning.

Unlike traditional supervised classification, complementary label learning (CLL) operates under a weak supervision framework, where each sample is annotated by excluding several incorrect labels, known as complementary labels (CLs). Despite reducing the labeling burden, CLL always suffers a decline in performance due to the weakened supervised information. To overcome such limitations, in this study, a multi-view fusion and self-adaptive label discovery based CLL method (MVSLDCLL) is proposed. The self-adaptive label discovery strategy leverages graph-based semi-supervised learning to capture the label distribution of each training sample as a convex combination of all its potential labels. The multi-view fusion module is designed to adapt to various views of feature representations. In specific, it minimizes the discrepancies of label projections between pairwise views, aligning with the consensus principle. Additionally, a straightforward mechanism inspired by a teamwork analogy is proposed to incorporate view-discrepancy for each sample. Experimental results demonstrate that MVSLDCLL learns more discriminative label distribution and achieves significantly higher accuracies compared to state-of-the-art CLL methods. Ablation study has also been performed to validate the effectiveness of both the self-adaptive label discovery strategy and the multi-view fusion module.

Optimal layer selection for latent data augmentation.

While data augmentation (DA) is generally applied to input data, several studies have reported that applying DA to hidden layers in neural networks, i.e., feature augmentation, can improve performance. However, in previous studies, the layers to which DA is applied have not been carefully considered, often being applied randomly and uniformly or only to a specific layer, leaving room for arbitrariness. Thus, in this study, we investigated the trends of suitable layers for applying DA in various experimental configurations, e.g., training from scratch, transfer learning, various dataset settings, and different models. In addition, to adjust the suitable layers for DA automatically, we propose the adaptive layer selection (AdaLASE) method, which updates the ratio to perform DA for each layer based on the gradient descent method during training. The experimental results obtained on several image classification datasets indicate that the proposed AdaLASE method altered the ratio as expected and achieved high overall test accuracy.

Accelerating multi-coil MR image reconstruction using weak supervision.

Deep-learning-based MR image reconstruction in settings where large fully sampled dataset collection is infeasible requires methods that effectively use both under-sampled and fully sampled datasets. This paper evaluates a weakly supervised, multi-coil, physics-guided approach to MR image reconstruction, leveraging both dataset types, to improve both the quality and robustness of reconstruction. A physics-guided end-to-end variational network (VarNet) is pretrained in a self-supervised manner using a 4 × under-sampled dataset following the self-supervised learning via data undersampling (SSDU) methodology. The pre-trained weights are transferred to another VarNet, which is fine-tuned using a smaller, fully sampled dataset by optimizing multi-scale structural similarity (MS-SSIM) loss in image space. The proposed methodology is compared with fully self-supervised and fully supervised training. Reconstruction quality improvements in SSIM, PSNR, and NRMSE when abundant training data is available (the high-data regime), and enhanced robustness when training data is scarce (the low-data regime) are demonstrated using weak supervision for knee and brain MR image reconstructions at 8 × and 10 × acceleration, respectively. Multi-coil physics-guided MR image reconstruction using both under-sampled and fully sampled datasets is achievable with transfer learning and fine-tuning. This methodology can provide improved reconstruction quality in the high-data regime and improved robustness in the low-data regime at high acceleration rates.

Robust and automatic beamstop shadow outlier rejection: combining crystallographic statistics with modern clustering under a semi-supervised learning strategy.

During the automatic processing of crystallographic diffraction experiments, beamstop shadows are often unaccounted for or only partially masked. As a result of this, outlier reflection intensities are integrated, which is a known issue. Traditional statistical diagnostics have only limited effectiveness in identifying these outliers, here termed Not-Excluded-unMasked-Outliers (NEMOs). The diagnostic tool AUSPEX allows visual inspection of NEMOs, where they form a typical pattern: clusters at the low-resolution end of the AUSPEX plots of intensities or amplitudes versus resolution. To automate NEMO detection, a new algorithm was developed by combining data statistics with a density-based clustering method. This approach demonstrates a promising performance in detecting NEMOs in merged data sets without disrupting existing data-reduction pipelines. Re-refinement results indicate that excluding the identified NEMOs can effectively enhance the quality of subsequent structure-determination steps. This method offers a prospective automated means to assess the efficacy of a beamstop mask, as well as highlighting the potential of modern pattern-recognition techniques for automating outlier exclusion during data processing, facilitating future adaptation to evolving experimental strategies.

A simulation data-driven semi-supervised framework based on MK-KNN graph and ESSGAT for bearing fault diagnosis.

Current supervised intelligent fault diagnosis relies on abundant labeled data. However, collecting and labeling data are typically both expensive and time-consuming. Fault diagnosis with unlabeled data remains a significant challenge. To address this issue, a simulation data-driven semi-supervised framework based on multi-kernel K-nearest neighbor (MK-KNN) and edge self-supervised graph attention network (ESSGAT) is proposed. The novel MK-KNN establishes the neighborhood relationships between simulation data and real data. The developed multi-kernel function mitigates the risks of overfitting and underfitting, thereby enhancing the robustness of the simulation-real graphs. The designed ESSGAT employs two forms of self-supervised attention to predict the presence of edges, increasing the weights of crucial neighboring nodes in the MK-KNN graph. The performance of the proposed method is evaluated using a public bearing dataset and a self-constructed dataset of high-speed train axle box bearings. The results show that the proposed method achieves better diagnostic performance compared with other state-of-the-art graph construction methods and graph convolutional networks.

Multi-label remote sensing classification with self-supervised gated multi-modal transformers.

Introduction: With the great success of Transformers in the field of machine learning, it is also gradually attracting widespread interest in the field of remote sensing (RS). However, the research in the field of remote sensing has been hampered by the lack of large labeled data sets and the inconsistency of data modes caused by the diversity of RS platforms. With the rise of self-supervised learning (SSL) algorithms in recent years, RS researchers began to pay attention to the application of "pre-training and fine-tuning" paradigm in RS. However, there are few researches on multi-modal data fusion in remote sensing field. Most of them choose to use only one of the modal data or simply splice multiple modal data roughly. Method: In order to study a more efficient multi-modal data fusion scheme, we propose a multi-modal fusion mechanism based on gated unit control (MGSViT). In this paper, we pretrain the ViT model based on BigEarthNet dataset by combining two commonly used SSL algorithms, and propose an intra-modal and inter-modal gated fusion unit for feature learning by combining multispectral (MS) and synthetic aperture radar (SAR). Our method can effectively combine different modal data to extract key feature information. Results and discussion: After fine-tuning and comparison experiments, we outperform the most advanced algorithms in all downstream classification tasks. The validity of our proposed method is verified.

Self-supervised in-domain representation learning for remote sensing image scene classification.

Transferring the ImageNet pre-trained weights to the various remote sensing tasks has produced acceptable results and reduced the need for labeled samples. However, the domain differences between ground imageries and remote sensing images cause the performance of such transfer learning to be limited. The difficulty of annotating remote sensing images is well-known as it requires domain experts and more time, whereas unlabeled data is readily available. Recently, self-supervised learning, which is a subset of unsupervised learning, emerged and significantly improved representation learning. Recent research has demonstrated that self-supervised learning methods capture visual features that are more discriminative and transferable than the supervised ImageNet weights. We are motivated by these facts to pre-train the in-domain representations of remote sensing imagery using contrastive self-supervised learning and transfer the learned features to other related remote sensing datasets. Specifically, we used the SimSiam algorithm to pre-train the in-domain knowledge of remote sensing datasets and then transferred the obtained weights to the other scene classification datasets. Thus, we have obtained state-of-the-art results on five land cover classification datasets with varying numbers of classes and spatial resolutions. In addition, by conducting appropriate experiments, including feature pre-training using datasets with different attributes, we have identified the most influential factors that make a dataset a good choice for obtaining in-domain features. We have transferred the features obtained by pre-training SimSiam on remote sensing datasets to various downstream tasks and used them as initial weights for fine-tuning. Moreover, we have linearly evaluated the obtained representations in cases where the number of samples per class is limited. Our experiments have demonstrated that using a higher-resolution dataset during the self-supervised pre-training stage results in learning more discriminative and general representations.

Rethinking deep clustering paradigms: Self-supervision is all you need.

The recent advances in deep clustering have been made possible by significant progress in self-supervised and pseudo-supervised learning. However, the trade-off between self-supervision and pseudo-supervision can give rise to three primary issues. The joint training causes Feature Randomness and Feature Drift, whereas the independent training causes Feature Randomness and Feature Twist. In essence, using pseudo-labels generates random and unreliable features. The combination of pseudo-supervision and self-supervision drifts the reliable clustering-oriented features. Moreover, moving from self-supervision to pseudo-supervision can twist the curved latent manifolds. This paper addresses the limitations of existing deep clustering paradigms concerning Feature Randomness, Feature Drift, and Feature Twist. We propose a new paradigm with a new strategy that replaces pseudo-supervision with a second round of self-supervision training. The new strategy makes the transition between instance-level self-supervision and neighborhood-level self-supervision smoother and less abrupt. Moreover, it prevents the drifting effect that is caused by the strong competition between instance-level self-supervision and clustering-level pseudo-supervision. Moreover, the absence of the pseudo-supervision prevents the risk of generating random features. With this novel approach, our paper introduces a Rethinking of the Deep Clustering Paradigms, denoted by R-DC. Our model is specifically designed to address three primary challenges encountered in Deep Clustering: Feature Randomness, Feature Drift, and Feature Twist. Experimental results conducted on six datasets have shown that the two-level self-supervision training yields substantial improvements, as evidenced by the results of the clustering and ablation study. Furthermore, experimental comparisons with nine state-of-the-art clustering models have clearly shown that our strategy leads to a significant enhancement in performance.

Reconstructing suspended sediment concentrations in the Mekong River Basin via semi-supervised-based deep neural networks.

The Mekong River Basin (MRB) is crucial for the livelihoods of over 60 million people across six Southeast Asian countries. Understanding long-term sediment changes is crucial for management and contingency plans, but the sediment concentration data in the MRB are extremely sporadic, making analysis challenging. This study focuses on reconstructing long-term suspended sediment concentration (SSC) data using a novel semi-supervised machine learning (ML) model. The key idea of this approach is to exploit abundant available hydroclimate data to reduce training overfitting rather than solely relying on sediment concentration data, thus enhancing the accuracy of the employed ML models. Extensive experiments on daily hydroclimate and SSC data obtained from 1979 to 2019 at the three main stations (i.e., Chiang Saen, Nong Khai, and Mukdahan) are conducted to demonstrate the superior performance of the proposed method compared to the state-of-the-art supervised techniques (i.e., Random Forest, XGBoost, CatBoost, MLP, CNN, and LSTM), and surpasses existing semi-supervised methods (i.e., CoReg, ⊓ Model, ICT, and Mean Teacher). This approach is the first semi-supervised method to reconstruct sediment data in the field and has the potential for broader application in other river systems.

Bridging vision and touch: advancing robotic interaction prediction with self-supervised multimodal learning.

Predicting the consequences of the agent's actions on its environment is a pivotal challenge in robotic learning, which plays a key role in developing higher cognitive skills for intelligent robots. While current methods have predominantly relied on vision and motion data to generate the predicted videos, more comprehensive sensory perception is required for complex physical interactions such as contact-rich manipulation or highly dynamic tasks. In this work, we investigate the interdependence between vision and tactile sensation in the scenario of dynamic robotic interaction. A multi-modal fusion mechanism is introduced to the action-conditioned video prediction model to forecast future scenes, which enriches the single-modality prototype with a compressed latent representation of multiple sensory inputs. Additionally, to accomplish the interactive setting, we built a robotic interaction system that is equipped with both web cameras and vision-based tactile sensors to collect the dataset of vision-tactile sequences and the corresponding robot action data. Finally, through a series of qualitative and quantitative comparative study of different prediction architecture and tasks, we present insightful analysis of the cross-modality influence between vision, tactile and action, revealing the asymmetrical impact that exists between the sensations when contributing to interpreting the environment information. This opens possibilities for more adaptive and efficient robotic control in complex environments, with implications for dexterous manipulation and human-robot interaction.

Weakly supervised learning for subcutaneous edema segmentation of abdominal CT using pseudo-labels and multi-stage nnU-Nets.

Volumetric assessment of edema due to anasarca can help monitor the progression of diseases such as kidney, liver or heart failure. The ability to measure edema non-invasively by automatic segmentation from abdominal CT scans may be of clinical importance. The current state-of-the-art method for edema segmentation using intensity priors is susceptible to false positives or under-segmentation errors. The application of modern supervised deep learning methods for 3D edema segmentation is limited due to challenges in manual annotation of edema. In the absence of accurate 3D annotations of edema, we propose a weakly supervised learning method that uses edema segmentations produced by intensity priors as pseudo-labels, along with pseudo-labels of muscle, subcutaneous and visceral adipose tissues for context, to produce more refined segmentations with demonstrably lower segmentation errors. The proposed method employs nnU-Nets in multiple stages to produce the final edema segmentation. The results demonstrate the potential of weakly supervised learning using edema and tissue pseudo-labels in improved quantification of edema for clinical applications.