Unlabeled Samples Research Articles

4hJCHOCH Spin Coupling in a Lewisx Trisaccharide as Evidence of Inter-Residue C-H···O Hydrogen Bonding in Aqueous Solution.

Prior studies of the solution conformation of the Lewisx (Lex) trisaccharide, αFuc-(1→3)[βGal-(1→4)]-βGlcNAc, suggest that nonclassical inter-residue C-H···O hydrogen bonding in aqueous solution contributes to the stabilization of its 3D structure and affects its biological properties. Experimental evidence for this hydrogen bond in aqueous solution has been reported in the form of a 4hJCHOCH NMR spin-coupling constant between C5'Fuc and H1″Gal measured by 2D NMR methods in unlabeled samples. A methyl glycoside of Lex (MeβLex) was prepared containing selective 13C-labeling at C5'Fuc, and the H1″Gal signal was examined in high-field 1H NMR spectra for evidence of splitting or line-broadening caused by the 13C at C5'Fuc. High-resolution 1H NMR spectra obtained at high field and at different temperatures using different FID processing parameters showed no resolved splitting of the H1″Gal signal or evidence of line-broadening. Spectral simulation showed that this splitting and/or line-broadening would be observable if the reported J-value (∼1.1 Hz) is correct. DFT calculations on MeβLex and a carbon analog (O5″Gal replaced by a CH2 group) gave very small and nearly identical calculated 4hJC5',H1″ values, suggesting that the coupling is essentially zero. DFT calculations also showed that an alternate inter-residue 3hJH5',H1″ is small. Based on NMR analyses and DFT calculations, we found that 4hJC5',H1″ in MeβLex has an upper limit of ∼0.4 Hz and that the value could be lower, possibly zero, calling into question its value as experimental proof of persistent nonclassical hydrogen bonding in aqueous solutions of MeβLex and related structures.

SOAMC: A Semi-Supervised Open-Set Recognition Algorithm for Automatic Modulation Classification

Traditional automatic modulation classification methods operate under the closed-set assumption, which proves to be impractical in real-world scenarios due to the diverse nature of wireless technologies and the dynamic characteristics of wireless propagation environments. Open-set environments introduce substantial technical challenges, particularly in terms of detection effectiveness and computational complexity. To address the limitations of modulation classification and recognition in open-set scenarios, this paper proposes a semi-supervised open-set recognition approach, termed SOAMC (Semi-Supervised Open-Set Automatic Modulation Classification). The primary objective of SOAMC is to accurately classify unknown modulation types, even when only a limited subset of samples is manually labeled. The proposed method consists of three key stages: (1) A signal recognition pre-training model is constructed using data augmentation and adaptive techniques to enhance robustness. (2) Feature extraction and embedding are performed via a specialized extraction network. (3) Label propagation is executed using a graph convolutional neural network (GCN) to efficiently annotate the unlabeled signal samples. Experimental results demonstrate that SOAMC significantly improves classification accuracy, particularly in challenging scenarios with limited amounts of labeled data and high signal similarity. These findings are critical for the practical identification of complex and diverse modulation signals in real-world wireless communication systems.

Semi-Supervised Online Continual Learning for 3D Object Detection in Mobile Robotics

Continual learning addresses the challenge of acquiring and retaining knowledge over time across multiple tasks and environments. Previous research primarily focuses on offline settings where models learn through increasing tasks from samples paired with ground truth annotations. In this work, we focus on an unsolved, challenging, yet practical scenario, specifically, the semi-supervised online continual learning in autonomous driving and mobile robotics. In our settings, models are tasked with learning new distributions from streaming unlabeled samples and performing 3D object detection as soon as the LiDAR point cloud arrives. Additionally, we conducted experiments on both the KITTI dataset, our newly built IUSL dataset and Canadian Adverse Driving Conditions (CADC) Dataset. The results indicate that our method achieves a balance between rapid adaptation and knowledge retention, showcasing its effectiveness in the dynamic and complex environment of autonomous driving and mobile robotics. The developed ROS packages and IUSL dataset will be publicly available at: https://github.com/npu-ius-lab/OCL3D.

Semi-Supervised Soft Computing for Ammonia Nitrogen Using a Self-Constructing Fuzzy Neural Network with an Active Learning Mechanism

Ammonia nitrogen (NH3-N) is a key water quality variable that is difficult to measure in the water treatment process. Data-driven soft computing is one of the effective approaches to address this issue. Since the detection cost of NH3-N is very expensive, a large number of NH3-N values are missing in the collected water quality dataset, that is, a large number of unlabeled data are obtained. To enhance the prediction accuracy of NH3-N, a semi-supervised soft computing method using a self-constructing fuzzy neural network with an active learning mechanism (SS-SCFNN-ALM) is proposed in this study. In the SS-SCFNN-ALM, firstly, to reduce the computational complexity of active learning, the kernel k-means clustering algorithm is utilized to cluster the labeled and unlabeled data, respectively. Then, the clusters with larger information values are selected from the unlabeled data using a distance metric criterion. Furthermore, to improve the quality of the selected samples, a Gaussian regression model is adopted to eliminate the redundant samples with large similarity from the selected clusters. Finally, the selected unlabeled samples are manually labeled, that is, the NH3-N values are added into the dataset. To realize the semi-supervised soft computing of the NH3-N concentration, the labeled dataset and the manually labeled samples are combined and sent to the developed SCFNN. The experimental results demonstrate that the test root mean square error (RMSE) and test accuracy of the proposed SS-SCFNN-ALM are 0.0638 and 86.31%, respectively, which are better than the SCFNN (without the active learning mechanism), MM, DFNN, SOFNN-HPS, and other comparison algorithms.

Inversion of biophysical parameters of potato based on an active learning pool-based sampling strategy

ABSTRACT Timely estimations of leaf chlorophyll content (LCC) and leaf area index (LAI) can provide critical information for potato field management. We employed a hybrid method that integrated machine learning with a radiative transfer model to estimate potato growth parameters. A look-up table was generated using the PROSAIL model, which was used as an unlabelled sample set. Measurements were taken from a potato field, and the data were labelled according to growth period and variety. Then, training samples for potato LCC, LAI, and canopy chlorophyll content (CCC) were selected from the simulated unlabelled sample set using the Euclidean distance-based diversity algorithm based on different labelled data sets. The training sample size required to accurately estimate the parameters varied considerably depending on the parameter, variety, and growth period, despite using the same labelled data set. Moreover, our results indicate that the growth period has a substantial impact on model accuracy and needs to be considered when constructing the labelled data set. The study results indicate that the hybrid method combined with the radiative transfer model and active learning can effectively select informative training samples from a data pool and improve the accuracy of potato parameter estimation, which provides a valid tool for accurately monitoring crop growth and growth health.

MisORFPred: A Novel Method to Mine Translatable sORFs in Plant Pri-miRNAs Using Enhanced Scalable k-mer and Dynamic Ensemble Voting Strategy.

The primary microRNAs (pri-miRNAs) have been observed to contain translatable small open reading frames (sORFs) that can encode peptides as an independent element. Relevant studies have proven that those of sORFs are of significance in regulating the expression of biological traits. The existing methods for predicting the coding potential of sORFs frequently overlook this data or categorize them as negative samples, impeding the identification of additional translatable sORFs in pri-miRNAs. In light of this, a novel method named misORFPred has been proposed. Specifically, an enhanced scalable k-mer (ESKmer) that simultaneously integrates the composition information within a sequence and distance information between sequences is designed to extract the nucleotide sequence features. After feature selection, the optimal features and several machine learning classifiers are combined to construct the ensemble model, where a newly devised dynamic ensemble voting strategy (DEVS) is proposed to dynamically adjust the weights of base classifiers and adaptively select the optimal base classifiers for each unlabeled sample. Cross-validation results suggest that ESKmer and DEVS are essential for this classification task and could boost model performance. Independent testing results indicate that misORFPred outperforms the state-of-the-art methods. Furthermore, we execute misORFPerd on the genomes of various plant species and perform a thorough analysis of the predicted outcomes. Taken together, misORFPred is a powerful tool for identifying the translatable sORFs in plant pri-miRNAs and can provide highly trusted candidates for subsequent biological experiments.

Fault diagnosis based on incomplete sensor variables with a hierarchical semi-supervised Gaussian mixture classifier

We propose a hierarchical semi-supervised variational semi-Bayesian Gaussian mixture classifier based on the partially incomplete and unlabeled samples for the fault diagnosis of mechanical and electrical systems. These systems are typically complex in structure and are monitored by multiple sensors simultaneously. Some sensor variables in the collected data may be incomplete due to sensor malfunctions or transmission errors. Additionally, labeling the data can be a time-consuming and labor-intensive task, resulting in many unlabeled samples. To address these challenges, the missing sensor variables and the unavailable labels are treated as hidden values and handled within the framework of the Expectation Maximization algorithm. We employ a semi-Bayesian technique with variational inference to estimate the model parameters. Specifically, we introduce prior distribution to the mean and covariance matrix to address the possible singularity of the empirical covariance matrix. The weighting coefficients are left without putting prior distribution so that their values can decay to zero, effectively removing redundant components of the mixture model. The factorized distribution is utilized to approximate the posterior distribution of the model parameters, as well as the missing labels and other latent variables. Numerical and real case studies are carried out to verify the effectiveness of the proposed technique.

Adaptive wavelet-VNet for single-sample test time adaptation in medical image segmentation.

In medical image segmentation, a domain gap often exists between training and testing datasets due to different scanners or imaging protocols, which leads to performance degradation in deep learning-based segmentation models. Given the high cost of manual labeling and the need for privacy protection, it is often challenging to annotate the testing (target) domain data for model fine-tuning or to collect data from different domains to train domain generalization models. Therefore, using only unlabeled target domain data for test-time adaptation (TTA) presents a more practical but challenging solution. To improve the segmentation accuracy of deep learning-based models on unseen datasets, and especially to enhance the efficiency and stability of TTA for individual samples from heterogeneous domains. In this study, we proposed to dynamically adapt a wavelet-VNet (WaVNet) to unseen target domains with a hybrid objective function, based on each unlabeled test sample during the test time. We embedded multiscale wavelet coefficients into a V-Net encoder and adaptively adjusted the spatial and spectral features according to the input, and the model parameters were optimized by three loss functions. We integrated a shape-aware loss to focus on the foreground segmentations, a Refine loss to correct the incomplete and noisy segmentations caused by domain shifts, and an entropy loss to promote the global consistency of the segmentations. We evaluated the proposed method on multidomain liver and prostate segmentation datasets to assess its advantages over other TTA methods. For the source domain model training of the liver dataset, we used 15 3D MR image samples for training and 5 for validation. Correspondingly, for the prostate dataset, we used 22 3D MR image samples for training and 7 for validation. In the target domain, we used a single 3D MR image sample for adaptation and testing. The total number of testing samples is 60 in the liver dataset (for 3 different domains) and 116 in the prostate dataset (for 6 different domains). The proposed method showed the highest segmentation accuracy among all methods, achieving a mean(±SD) Dice coefficient (DSC) of 78.10±5.23% and a mean 95th Hausdorff distance (HD95) of 15.52±5.84mm on the liver dataset; and a mean DSC of 80.02±3.89% and a mean HD95 of 9.18±3.47mm on the prostate dataset. The DSC is 11.67% (in absolute terms) and 15.27% higher than that of the baseline (no adaptation) method, for the liver and the prostate datasets, respectively. The proposed adaptive WaVNet enhanced the image segmentation accuracy from unseen domains during the test time via unsupervised learning and multi-objective optimization. It can benefit clinical applications where data are scarce or with changing data distributions, including online adaptive radiotherapy. The code will be released at: https://github.com/sanny1226/WaVNet.

Learning protein language contrastive models with multi-knowledge representation

Protein representation learning plays a crucial role in obtaining a comprehensive understanding of biological regulatory mechanisms and in developing proteins and drugs for therapeutic purposes. However, labeled proteins, such as sequenced and functionally annotated data, are incomplete and few. Thus, contrastive learning has emerged as the preferred technique for learning meaningful representations from unlabeled data samples. In addition, at present, natural proteins cannot be fully described by extracting protein knowledge from a single domain. Therefore, Pro-CoRL, a protein contrastive models framework based on multi-knowledge representation learning, was proposed in this study. In particular, Pro-CoRL smooths the objective function using convex approximation, thereby improving the stability of training. Extensive experiments on predicting protein–protein interaction types and clustering protein families have confirmed the high accuracy and robustness of Pro-CoRL.

A novel self-training framework for semi-supervised soft sensor modeling based on indeterminate variational autoencoder

In modern industrial processes, the high acquisition cost of labeled data can lead to a large number of unlabeled samples, which greatly impacts the accuracy of traditional soft sensor models. To this end, this paper proposes a novel semi-supervised soft sensor framework that can fully utilize the unlabeled data to expand the original labeled data, and ultimately improve the prediction accuracy. Specifically, an indeterminate variational autoencoder (IVAE) is first proposed to obtain the pseudo-labels of unlabeled data as well as their uncertainty. On this basis, the IVAE-based self-training (ST-IVAE) framework is further naturally proposed to expand the original small labeled dataset through continuous circulation. Among them, a variance-based oversampling (VOS) strategy is introduced to better utilize the pseudo-label uncertainty. By determining similar sample sets through the comparison of Kullback-Leibler (KL) divergence obtained by the proposed IVAE model, each sample can be independently modeled for prediction. The effectiveness of the proposed semi-supervised framework is verified on two real industrial processes and the realization results illustrate that the ST-IVAE framework can still predict well even in the presence of missing input data compared to state-of-the-art methodologies in addressing semi-supervised soft sensing challenges.

Active Learning for Deep Visual Tracking.

Convolutional neural networks (CNNs) have been successfully applied to the single target tracking task in recent years. Generally, training a deep CNN model requires numerous labeled training samples, and the number and quality of these samples directly affect the representational capability of the trained model. However, this approach is restrictive in practice, because manually labeling such a large number of training samples is time-consuming and prohibitively expensive. In this article, we propose an active learning method for deep visual tracking, which selects and annotates the unlabeled samples to train the deep CNN model. Under the guidance of active learning, the tracker based on the trained deep CNN model can achieve competitive tracking performance while reducing the labeling cost. More specifically, to ensure the diversity of selected samples, we propose an active learning method based on multiframe collaboration to select those training samples that should be and need to be annotated. Meanwhile, considering the representativeness of these selected samples, we adopt a nearest-neighbor discrimination method based on the average nearest-neighbor distance to screen isolated samples and low-quality samples. Therefore, the training samples' subset selected based on our method requires only a given budget to maintain the diversity and representativeness of the entire sample set. Furthermore, we adopt a Tversky loss to improve the bounding box estimation of our tracker, which can ensure that the tracker achieves more accurate target states. Extensive experimental results confirm that our active-learning-based tracker (ALT) achieves competitive tracking accuracy and speed compared with state-of-the-art trackers on the seven most challenging evaluation benchmarks. Project website: https://sites.google.com/view/altrack/.

Perspective-assisted prototype-based learning for semi-supervised crowd counting

To alleviate the burden of labeling data to train crowd counting models, we propose a prototype-based learning approach for semi-supervised crowd counting with an embeded understanding of perspective. Our key idea is that image patches with the same density of people are likely to exhibit coherent appearance changes under similar perspective distortion, but differ significantly under varying distortions. Motivated by this observation, we construct multiple prototypes for each density level to capture variations in perspective. For labeled data, the prototype-based learning assists the regression task by regularizing the feature space and modeling the relationships within and across different density levels. For unlabeled data, the learnt perspective-embedded prototypes enhance differentiation between samples of the same density levels, allowing for a more nuanced assessment of the predictions. By incorporating regression results, we categorize unlabeled samples as reliable or unreliable, applying tailored consistency learning strategies to enhance model accuracy and generalization. Since the perspective information is often unavailable, we propose a novel pseudo-label assigner based on perspective self-organization which requires no additional annotations and assigns image regions to distinct spatial density groups, which mainly reflect the differences in average density among regions. Extensive experiments on four crowd counting benchmarks demonstrate the effectiveness of our approach.

Self-adjusted graph based semi-supervised embedded feature selection

Graph-based semi-supervised feature selection has aroused continuous attention in processing high-dimensional data with most unlabeled and fewer data samples. Many graph-based models perform on a pre-defined graph, which is separated from the procedure of feature selection, making the model hard to select the discriminative features. To address this issue, we exploit a self-adjusted graph for semi-supervised embedded feature selection method (SAGFS), which learns an optimal sparse similarity graph to replace the pre-defined graph to alleviate the effect of data noise. SAGFS allows the learned graph itself to be adjusted according to the local geometric structure of the data and the procedure of selecting features to select the most representative features. Besides that, we introduce l2,p\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$l_{2,p}$$\\end{document}-norm to constrain the projection matrix for efficient feature selection. An efficient alternating optimization algorithm is presented, together with analyses on its convergence. Systematical experiments on several publicly datasets are performed to analyze the proposed model from several aspects, and demonstrate that our approaches outperform other comparison methods.

Lassi onk: a system framework to annotate and classify vehicular honk from road traffic.

Some recent studies highlight that vehicular traffic and honking contribute to more than 50% of noise pollution in urban or sub-urban areas in developing countries, including Indian cities. Frequent honking has an adverse effect on health and hampers road safety, the environment, etc. Therefore, recognizing the various vehicle honks and classifying the honk of different vehicles can provide good insights into environmental noise pollution. Moreover, classifying honks based on vehicle types allows for the inference of contextual information of a location, area, or traffic. So far, the researchers have done outdoor sound classification and honk detection, where vehicular honks are collected in a controlled environment or in theabsence of ambient noise. Such classification models fail to classify honk based on vehicle types. Therefore, it becomes imperative to design a system that can detect and classify honks of different types of vehicles to infer some contextual information. This paper presents a novel framework lassi onk that performs raw vehicular honk sensing, data labeling, and classifies the honk into three major groups, i.e., light-weight vehicles, medium-weight vehicles, and heavy-weight vehicles. Raw audio samples of different vehicular honking are collected based on spatio-temporal characteristics and converted them into spectrogram images. A deep learning-based multi-label autoencoder model (MAE) is proposed for automated labeling of the unlabeled data samples, which provides 97.64% accuracy in contrast to existing deep learning-based data labeling methods. Further, various pre-trained models, namely Inception V3, ResNet50, MobileNet, and ShuffleNet are used and proposed an Ensembled Transfer Learning model (EnTL) for vehicle honks classification and performed comparative analysis. Results reveal that EnTL exhibits the best performance compared to pre-trained models and achieves 96.72% accuracy in our dataset. In addition, context of a location is identified based on these classified honk signatures in a city.

High-Sensitivity Analysis of Native Bacterial Biofilms Using Dynamic Nuclear Polarization-Enhanced Solid-State NMR.

Bacterial biofilms cause persistent infections that are difficult to treat and contribute greatly to antimicrobial resistance. However, high-resolution structural information on native bacterial biofilms remain very limited. This limitation is primarily due to methodological constraints associated with analyzing complex native samples. Although solid-state NMR (ssNMR) is a promising method in this regard, its conventional applications typically suffer from sensitivity limitations, particularly for unlabeled native samples. Through the use of Dynamic Nuclear Polarization (DNP), we applied sensitivity enhanced ssNMR to characterize native Pseudomonas fluorescens colony biofilms. The increased ssNMR sensitivity by DNP enabled ultrafast structural characterization of the biofilm samples without isotope-labelling, and chemical or physical modification. We collected 1D 13 C and 15 N, and 2D 1 H- 13 C, 1 H- 15 N and 13 C- 13 C ssNMR spectra within seconds/minutes or hours, respectively which enabled us to identify biofilm components as polysaccharides, proteins, and eDNA effectively. This study represents the first application of ultrasensitive DNP ssNMR to characterize a native bacterial biofilm and expands the technical scope of ssNMR towards obtaining insights into the composition and structure of a wide array of in vitro and ex vivo biofilm applications. Such versatility should greatly boost efforts to develop structure-guided approaches for combating infections caused by biofilm-forming microbes.

FireDA: A Domain Adaptation-Based Method for Forest Fire Recognition with Limited Labeled Scenarios

Vision-based forest fire detection systems have significantly advanced through Deep Learning (DL) applications. However, DL-based models typically require large-scale labeled datasets for effective training, where the quality of data annotation is crucial to their performance. To address challenges related to the quality and quantity of labeling, a domain adaptation-based approach called FireDA is proposed for forest fire recognition in scenarios with limited labels. Domain adaptation, a subfield of transfer learning, facilitates the transfer of knowledge from a labeled source domain to an unlabeled target domain. The construction of the source domain FBD is initiated, which includes three common fire scenarios: forest (F), brightness (B), and darkness (D), utilizing publicly available labeled data. Subsequently, a novel algorithm called Neighborhood Aggregation-based 2-Stage Domain Adaptation (NA2SDA) is proposed. This method integrates feature distribution alignment with target domain Proxy Classification Loss (PCL), leveraging a neighborhood aggregation mechanism and a memory bank designed for the unlabeled samples in the target domain. This mechanism calibrates the source classifier and generates more accurate pseudo-labels for the unlabeled sample. Consequently, based on these pseudo-labels, the Local Maximum Mean Discrepancy (LMMD) and the Proxy Classification Loss (PCL) are computed. To validate the efficacy of the proposed method, the publicly available forest fire dataset, FLAME, is employed as the target domain for constructing a transfer learning task. The results demonstrate that our method achieves performance comparable to the supervised Convolutional Neural Network (CNN)-based state-of-the-art (SOTA) method, without requiring access to labels from the FLAME training set. Therefore, our study presents a viable solution for forest fire recognition in scenarios with limited labeling and establishes a high-accuracy benchmark for future research.

Semi-supervised recognition for artificial intelligence assisted pathology image diagnosis

The analysis and interpretation of cytopathological images are crucial in modern medical diagnostics. However, manually locating and identifying relevant cells from the vast amount of image data can be a daunting task. This challenge is particularly pronounced in developing countries where there may be a shortage of medical expertise to handle such tasks. The challenge of acquiring large amounts of high-quality labelled data remains, many researchers have begun to use semi-supervised learning methods to learn from unlabeled data. Although current semi-supervised learning models partially solve the issue of limited labelled data, they are inefficient in exploiting unlabeled samples. To address this, we introduce a new AI-assisted semi-supervised scheme, the Reliable-Unlabeled Semi-Supervised Segmentation (RU3S) model. This model integrates the ResUNet-SE-ASPP-Attention (RSAA) model, which includes the Squeeze-and-Excitation (SE) network, Atrous Spatial Pyramid Pooling (ASPP) structure, Attention module, and ResUNet architecture. Our model leverages unlabeled data effectively, improving accuracy significantly. A novel confidence filtering strategy is introduced to make better use of unlabeled samples, addressing the scarcity of labelled data. Experimental results show a 2.0% improvement in mIoU accuracy over the current state-of-the-art semi-supervised segmentation model ST, demonstrating our approach’s effectiveness in solving this medical problem.

Structural tensor and frequency guided semi-supervised segmentation for medical images.

The method of semi-supervised semantic segmentation entails training with a limited number of labeled samples alongside many unlabeled samples, aiming to reduce dependence on pixel-level annotations. Most semi-supervised semantic segmentation methods primarily focus on sample augmentation in spatial dimensions to reduce the shortage of labeled samples. These methods tend to ignore the structural information of objects. In addition, frequency-domain information also supplies another perspective to evaluate information from images, which includes different properties compared to the spatialdomain. In this study, we attempt to answer these two questions: (1) is it helpful to provide structural information of objects in semi-supervised semantic segmentation tasks for medical images? (2) is it more effective to evaluate the segmentation performance in the frequency domain compared to the spatial domain for semi-supervised medical image segmentation? Therefore, we seek to introduce structural and frequency information to improve the performance of semi-supervised semantic segmentation for medicalimages. We present a novel structural tensor loss (STL) to guide feature learning on the spatial domain for semi-supervised semantic segmentation. Specifically, STL utilizes the structural information encoded in the tensors to enforce the consistency of objects across spatial regions, thereby promoting more robust and accurate feature extraction. Additionally, we proposed a frequency-domain alignment loss (FAL) to enable the neural networks to learn frequency-domain information across different augmented samples. It leverages the inherent patterns present in frequency-domain representations to guide the network in capturing and aligning features across diverse augmentation variations, thereby enhancing the model's robustness for the inputting variations. We conduct our experiments on three benchmark datasets, which include MRI (ACDC) for cardiac, CT (Synapse) for abdomen organs, and ultrasound image (BUSI) for breast lesion segmentation. The experimental results demonstrate that our method outperforms state-of-the-art semi-supervised approaches regarding the Dice similaritycoefficient. We find the proposed approach could improve the final performance of the semi-supervised medical image segmentation task. It will help reduce the need for medical image labels. Our code will are available at https://github.com/apple1986/STLFAL.

Graph neural pre-training based drug-target affinity prediction.

Computational drug-target affinity prediction has the potential to accelerate drug discovery. Currently, pre-training models have achieved significant success in various fields due to their ability to train the model using vast amounts of unlabeled data. However, given the scarcity of drug-target interaction data, pre-training models can only be trained separately on drug and target data, resulting in features that are insufficient for drug-target affinity prediction. To address this issue, in this paper, we design a graph neural pre-training-based drug-target affinity prediction method (GNPDTA). This approach comprises three stages. In the first stage, two pre-training models are utilized to extract low-level features from drug atom graphs and target residue graphs, leveraging a large number of unlabeled training samples. In the second stage, two 2D convolutional neural networks are employed to combine the extracted drug atom features and target residue features into high-level representations of drugs and targets. Finally, in the third stage, a predictor is used to predict the drug-target affinity. This approach fully utilizes both unlabeled and labeled training samples, enhancing the effectiveness of pre-training models for drug-target affinity prediction. In our experiments, GNPDTA outperforms other deep learning methods, validating the efficacy of our approach.

Learning Unsupervised Cross-Domain Model for TIR Target Tracking

The limited availability of thermal infrared (TIR) training samples leads to suboptimal target representation by convolutional feature extraction networks, which adversely impacts the accuracy of TIR target tracking methods. To address this issue, we propose an unsupervised cross-domain model (UCDT) for TIR tracking. Our approach leverages labeled training samples from the RGB domain (source domain) to train a general feature extraction network. We then employ a cross-domain model to adapt this network for effective target feature extraction in the TIR domain (target domain). This cross-domain strategy addresses the challenge of limited TIR training samples effectively. Additionally, we utilize an unsupervised learning technique to generate pseudo-labels for unlabeled training samples in the source domain, which helps overcome the limitations imposed by the scarcity of annotated training data. Extensive experiments demonstrate that our UCDT tracking method outperforms existing tracking approaches on the PTB-TIR and LSOTB-TIR benchmarks.