Semi-supervised multi-label learning (SSMLL) aims to address the challenge of limited labeled data in multi-label learning (MLL) by leveraging unlabeled data to improve the model’s performance. While pseudo-labeling has become a dominant strategy in SSMLL, most existing methods assign equal weights to all pseudo-labels regardless of their quality, which can amplify the impact of noisy or uncertain predictions and degrade the overall performance. In this paper, we theoretically verify that the optimal weight for a pseudo-label should reflect its correctness likelihood. Empirically, we observe that on the same dataset, the correctness likelihood distribution of unlabeled data remains stable, even as the number of labeled training samples varies. Building on this insight, we propose Distribution-Calibrated Pseudo-labeling (DiCaP), a correctness-aware framework that estimates posterior precision to calibrate pseudo-label weights. We further introduce a dual-thresholding mechanism to separate confident and ambiguous regions: confident samples are pseudo-labeled and weighted accordingly, while ambiguous ones are explored by unsupervised contrastive learning. Experiments conducted on multiple benchmark datasets verify that our method achieves consistent improvements, surpassing state-of-the-art methods by up to 4.27%.

This paper studies the long-tailed semi-supervised learning (LTSSL) with distribution mismatch, where the class distribution of the labeled training data follows a long-tailed distribution and mismatches with that of the unlabeled training data. Most existing methods introduce auxiliary classifiers (experts) to model various unlabeled data distributions and produce pseudo-labels, but the expertises of various experts are not fully utilized. We observe that different experts are good at predicting different intervals of samples, e.g., long-tailed expert is skilled in samples located in the head interval and uniform expert excels in samples located in the medium interval. Therefore, we propose a dynamic expert assignment module that can estimate the class membership (i.e., head, medium, or tail class) of samples, and dynamically assigns suitable expert to each sample based on the estimated membership to produce high-quality pseudo-label in the training phase and produce prediction in the testing phase. We also theoretically reveal that integrating different experts' strengths will lead to a smaller generalization error bound. Moreover, we find that the deeper features are more biased toward the head class but with more discriminative ability, while the shallower features are less biased but also with less discriminative ability. We, therefore, propose a multi-depth feature fusion module to utilize different depth features to mitigate the model bias. Our method demonstrates its effectiveness through comprehensive experiments on the CIFAR-10-LT, STL-10-LT, and SVHN-LT datasets across various settings.

Current long-tailed semi-supervised learning methods assume that labeled data exhibit a long-tailed distribution, and unlabeled data adhere to a typical predefined distribution (i.e., long-tailed, uniform, or inverse long-tailed). However, the distribution of the unlabeled data is generally unknown and may follow an arbitrary distribution. To tackle this challenge, we propose a Controllable Pseudo-label Generation (CPG) framework, expanding the labeled dataset with the progressively identified reliable pseudo-labels from the unlabeled dataset and training the model on the updated labeled dataset with a known distribution, making it unaffected by the unlabeled data distribution. Specifically, CPG operates through a controllable self-reinforcing optimization cycle: (i) at each training step, our dynamic controllable filtering mechanism selectively incorporates reliable pseudo-labels from the unlabeled dataset into the labeled dataset, ensuring that the updated labeled dataset follows a known distribution; (ii) we then construct a Bayes-optimal classifier using logit adjustment based on the updated labeled data distribution; (iii) this improved classifier subsequently helps identify more reliable pseudo-labels in the next training step. We further theoretically prove that this optimization cycle can significantly reduce the generalization error under some conditions. Additionally, we propose a class-aware adaptive augmentation module to further improve the representation of minority classes, and an auxiliary branch to maximize data utilization by leveraging all labeled and unlabeled samples. Comprehensive evaluations on various commonly used benchmark datasets show that CPG achieves consistent improvements, surpassing state-of-the-art methods by up to **15.97\%** in accuracy. The code is available at https://github.com/yaxinhou/CPG.

Neck lymph node detection is crucial for early cancer metastasis detection and treatment, influencing treatment success and patient survival rates. It also aids in disease staging, monitoring, and treatment selection. It requires the expertise of professional senior radiologists, as the accuracy of current automated detection methods is not sufficiently high. In this study, the neck lymph node detection network (LymoNet) based on YOLOv8 is proposed to detect and classify normal, inflammatory, and metastatic neck lymph nodes from ultrasound images. The advanced attention mechanism modules are utilized to enhance performance of the model, including the Coordinate Attention (CA) which helps the network focus on learning key features in the images, and the Multi-Head Self-Attention (MHSA) which captures global information at different scales. Meanwhile, the medical knowledge embedding which introduces prior knowledge from the medical domain is used to improve the classification performance. By integrating these elements, the YOLOv8 network can achieve better performance in neck lymph node detection tasks. Finally, LymoNet surpassed the benchmark model YOLOv8 by 6.6% in the [email protected], achieving the state-of-the-art (SOTA). This model provides a promising solution for automated neck lymph node detection in clinical environments. The proposed methods can also serve as a reference for applying deep learning algorithms in other fields. The source codes, trained weights, and validation data are available on GitHub.

Data scarcity is a very common real-world problem that poses a major challenge to data-driven analytics. Although a lot of data-balancing approaches have been proposed to mitigate this problem, they may drop some useful information or fall into the overfitting problem. Generative Adversarial Network (GAN) based data synthesis methods can alleviate such a problem but lack of quality control over the generated samples. Moreover, the latent associations between the attribute set and the class labels in a relational data cannot be easily captured by a vanilla GAN. In light of this, we introduce an end-to-end self-training scheme (namely, Quality-Aware Self-Training) for rare relational data synthesis, which generates labeled synthetic data via pseudo labeling on GAN-based synthesis. We design a semantic pseudo labeling module to first control the quality of the generated features/samples, then calibrate their semantic labels via a classifier committee consisting of multiple pre-trained shallow classifiers. The high-confident generated samples with calibrated pseudo labels are then fed into a semantic classification network as augmented samples for self-training. We conduct extensive experiments on 20 benchmark datasets of different domains, including 14 industrial datasets. The results show that our method significantly outperforms state-of-the-art methods, including two recent GAN-based data synthesis schemes. Codes are available at https://github.com/yaxinhou/QAST.

Due to the constraint of minimum antenna area, 2-D ambiguity resolving is a challenging task in ship focusing of single-channel SAR mounted on high-speed maneuvering platforms. In order to accommodate the issues, a ship focusing and positioning algorithm based on 2-D ambiguity resolving is proposed. First, we analyze the constraint of minimum antenna area and the distribution of the 2-D ambiguity area. An optimal-PRF SAR concept is proposed, where the signal ambiguity is evenly distributed to the range and azimuth directions. In this concept, the timing sequence of the orthogonal phase-coded waveform is designed to make the target echoes in different regions orthogonal. Then, an orthogonal matching filter is used to suppress the signal energy in range ambiguity regions. Aiming at the target defocus and position shift caused by the Doppler ambiguity, we propose an azimuth ambiguity resolving method. The Doppler ambiguity number can be estimated by residual envelope inclination. Subsequently, the target can be relocated and accurately focused by the estimated Doppler parameters. After the operation of each target is completed, the focusing SAR image of the whole scene can be obtained. Finally, simulation results and real data processing are presented to validate the proposed algorithm.