Deep learning-based automatic modulation recognition (AMR) models are highly susceptible to adversarial samples. Although adversarial training offers a promising defense, most existing methods rely on fixed-intensity attacks and lack mechanisms to maintain stable feature representations, limiting recognition performance under diverse adversarial conditions. To address these issues, we propose S3AT, an adversarial training framework that integrates Self-paced learning, Self-distillation, and Self-finetuning. In the self-paced phase, a progressive, difficulty-aware training paradigm adaptively schedules adversarial strength to help the model gradually adapt from weak to strong attacks. An exponential moving average (EMA) teacher provides continuous self-distillation throughout training to stabilize feature learning and mitigate forgetting of early knowledge. In the self-finetuning phase, historical checkpoints serve as auxiliary teachers, complementing the self-distillation and enhancing robustness across varying intensities of adversarial attacks. Experiments on public and self-collected datasets confirm that S3AT outperforms baseline methods under both glass-box and closed-box attacks. Its compatibility with external defenses further highlights its flexibility for real-world electromagnetic applications. Code is available at https://gitee.com/wwenyuu/s3-at

Change detection (CD) in remote sensing images plays a vital role in dynamic Earth measurement. However, the scarcity of high-resolution, comprehensive open-source datasets and the difficulty in achieving robust performance across varying change types remain major challenges. To address these issues, we introduce JL1-CD, a large-scale, submeter CD dataset consisting of 5000 image pairs. We further propose a novel origin-partition (O-P) strategy and integrate it into a multiteacher knowledge distillation (MTKD) framework to enhance CD performance. The O-P strategy partitions the training set by change area ratio (CAR) and trains specialized teacher models on each subset. The MTKD framework then distills complementary knowledge from these teachers into a single student model, enabling improved detection results across diverse CAR scenarios without additional inference cost. Our MTKD approach demonstrated strong performance in the 2024 "Jilin-1" Cup challenge, ranking first in the preliminary and second in the final rounds. Extensive experiments on the JL1-CD and SYSU-CD datasets show that the MTKD framework consistently improves the performance of CD models with various network architectures and parameter sizes, establishing new state-of-the-art results. Code and dataset are available at https://github.com/circleLZY/MTKD-CD

Classifier-Free Guidance (CFG) is widely used to improve conditional fidelity in diffusion models, but its impact on sampling dynamics remains poorly understood. Prior studies, often restricted to unimodal conditional distributions or simplified cases, provide only a partial picture. We analyze CFG under multimodal conditionals and show that the sampling process unfolds in three successive stages. In the Direction Shift stage, guidance accelerates movement toward the weighted mean, introducing initialization bias and norm growth. In the Mode Separation stage, local dynamics remain largely neutral, but the inherited bias suppresses weaker modes, reducing global diversity. In the Concentration stage, guidance amplifies within-mode contraction, diminishing fine-grained variability. This unified view explains a widely observed phenomenon: stronger guidance improves semantic alignment but inevitably reduces diversity. Experiments support these predictions, showing that early strong guidance erodes global diversity, while late strong guidance suppresses fine-grained variation. Moreover, our theory naturally suggests a time-varying guidance schedule, and empirical results confirm that it consistently improves both quality and diversity.

Few-step distilled diffusion models generate high-quality images quickly, but often lose per-prompt diversity, producing near-identical samples across random seeds. Optimizing the initial noise at inference time offers an appealing way to recover this diversity, yet existing methods directly update the initial noise in an unconstrained Euclidean space, ignoring both the geometry of the Gaussian prior and the model's sensitivity to noise frequencies. They therefore introduce auxiliary quality-control objectives to maintain generation fidelity, adding compute and weighting hyperparameters while still requiring conservative updates to prevent degradation. In this work, we propose MoNO, a training-free method that performs Manifold-constrained Noise Optimization on a low-dimensional, quality-stabilizing noise manifold. MoNO sequentially optimizes each new initial noise so that its predicted visual feature complements previous generations, while Riemannian updates on an affine low-frequency sphere preserve prior likelihood and fix unstable high-frequency components by construction. This enables large geodesic steps, removes the need for auxiliary quality-control objectives, and converges in far fewer iterations than prior noise-optimization methods. Experiments with multiple distilled text-to-image diffusion models show that MoNO consistently improves per-prompt diversity while maintaining image quality.

The Normalized Difference Vegetation Index (NDVI) is one of the most widely used remote sensing indicators for assessing vegetation status, with applications across a broad range of studies including phenology, ecology, and hydrology. However, NDVI data are usually degraded by clouds, snow cover, and other factors. In this study, we propose a novel spatiotemporal method-the Time-varying Graph Signal Reconstruction (TGSR) method-for the reconstruction of low-quality NDVI data. This method regards spatiotemporal NDVI data as time-varying graph signals and reconstructs them by maximizing the differential smoothness of the signal. We applied the TGSR method to sample patches (128 & times; 128 pixels) from MODIS/Terra Vegetation Indices 16-Day L3 Global 500 m Grid products across East Asia, and quantitatively compared the results with those from two benchmark methods: the Savitzky-Golay (SG) filter method and the Spatiotemporal Tensor (ST-Tensor) method. Results demonstrate that the TGSR method outperforms both benchmark methods in capturing temporal dynamics and preserving spatial structure of NDVI data, achieving an average root mean square error of 0.0302, compared to 0.0354 for SG and 0.0331 for ST-Tensor. The TGSR method exhibits no bias toward noise type-whether positively biased, negatively biased, or missing data-and shows strong robustness against varying noise intensities, performing well even with only 15% reliable data. Moreover, the TGSR method effectively addresses challenges related to spatiotemporal continuous gaps and land cover changes. The present findings highlight the strong potential of the TGSR method for generating high-quality NDVI datasets and its applicability to similar remote sensing time-series products.

The distortion–perception (D–P) tradeoff is a fundamental phenomenon of Bayesian inverse problems, which characterizes the inherent tension between distortion performance and perceptual quality. Enabling flexible traversal of the D-P tradeoff at inference time is crucial for practical applications. Despite the recent success of diffusion models in zero-shot inverse problem solving, efficient and principled strategies for D-P traversal in diffusion-based inverse algorithms remain inadequately characterized. In this paper, we propose a stage-wise framework for realizing D-P traversal using a single diffusion model in zero-shot inverse problems. Our proposed method, termed MAP-RPS, starts with an MAP estimation stage that approximates the MMSE solution and provides a low-distortion initialization, followed by a re-noised posterior sampling stage that progressively improves perceptual quality. We provide theoretical analyses for both stages, establishing the validity and effectiveness of the proposed design. Furthermore, we extend MAP-RPS to the latent space, yielding LMAP-RPS, which enjoys broader applicability by leveraging large-scale pre-trained latent diffusion backbones. Extensive experiments demonstrate that MAP-RPS and LMAP-RPS enable more effective D-P traversal on various tasks, while also exhibiting strong performance as efficient solvers for real-world inverse problems.

Remote sensing change detection (CD) traditionally focuses on pixel-level binary segmentation, which identifies where changes occur but neither what nor why. To bridge this semantic gap, we introduce JL1-CC QA, a multi-task benchmark that extends the JL1-CD dataset with two complementary annotation layers: change captioning (CC) and change question answering (QA). Built upon 5,000 bi-temporal image pairs acquired by the Jilin-1 satellite at 0.5-0.75m ground sample distance, the benchmark comprises: (i) JL1-CC, providing 17,021 quality-verified captions that describe diverse land-cover transformations; and (ii) JL1-QA, offering 20,060 question-answer pairs across eight question types, enabling fine-grained, interactive interrogation of surface changes. All annotations are produced via a three-stage pipeline consisting of multi-modal large language model (LLM) generation, vision-grounded LLM judging, and human expert verification. We hope that JL1-CC QA, as a benchmark unifying binary change masks, change captions, and change-oriented QA over the same image set, will serve as a valuable resource for the community to advance multi-task change understanding in remote sensing. The dataset is available at https://github.com/circleLZY/JL1-CD.

Token-based time series large language models (TS-LLMs) have emerged as a promising direction for time series analysis and reasoning. However, prior studies largely overlook the inherent continuity and ordinality of time series tokens, which substantially limits model performance. In this paper, we argue that preserving these properties in time series token embeddings is crucial for the effectiveness of token-based TS-LLMs. To this end, we propose COM (Continuity and Ordinality Matter), a continuity- and ordinality-aware strategy that integrates geometric constraints into both the initialization and training stages. Empirical results on multiple time series analysis benchmarks demonstrate that COM consistently improves the performance of token-based TS-LLMs, achieving competitive results and strong generalizability. Code is available at https://anonymous.4open.science/r/COM .

Time series data are central to domains such as finance, healthcare, and cloud computing, yet existing benchmarks for evaluating various large language models (LLMs) on temporal tasks remain scattered and unsystematic. To bridge this gap, we introduce MMTS-BENCH, a comprehensive multimodal benchmark built upon a hierarchical taxonomy of time-series tasks, spanning structural awareness, feature analysis, temporal reasoning, sequence matching and cross-modal alignment. MMTS-BENCH comprises 2,424 time series question answering (TSQA) pairs across 4 subsets: Base, InWild, Match, and Align, generated through a progressive real-world QA framework and modular synthetic data construction. We conduct extensive evaluations on closed-source, open-source LLMs and existing time series adapted large language models (TS-LLMs), revealing that: (1) TS-LLMs significantly lag behind general-purpose LLMs in cross-domain generalization, (2) LLMs show weaknesses in local tasks compared to global tasks, (3) chain-of-thought (CoT) reasoning and multimodal integration substantially improve performance, and (4) the dominant factor in existing TS-LLMs remains the backbone network capability rather than the time series encoder design. MMTS-BENCH not only provides a rigorous evaluation framework but also offers clear directions for advancing LLMs toward robust, interpretable, and generalizable time-series reasoning.

Diffusion models deliver state-of-the-art generative performance across diverse modalities but remain computationally expensive due to their inherently iterative sampling process. Existing training-free acceleration methods typically improve numerical solvers for the reverse-time ODE, yet their effectiveness is fundamentally constrained by the inefficiency of the underlying sampling trajectories. We propose A-FloPS (Adaptive Flow Path Sampler), a principled, training-free framework that reparameterizes the sampling trajectory of any pre-trained diffusion model into a flow-matching form and augments it with an adaptive velocity decomposition. The reparameterization analytically maps diffusion scores to flow-compatible velocities, yielding integration-friendly trajectories without retraining. The adaptive mechanism further factorizes the velocity field into a linear drift term and a residual component whose temporal variation is actively suppressed, restoring the accuracy benefits of high-order integration even in extremely low-NFE regimes. Extensive experiments on conditional image generation and text-to-image synthesis show that A-FloPS consistently outperforms state-of-the-art training-free samplers in both sample quality and efficiency. Notably, with as few as 5 function evaluations, A-FloPS achieves substantially lower FID and generates sharper, more coherent images. The adaptive mechanism also improves native flow-based generative models, underscoring its generality. These results position A-FloPS as a versatile and effective solution for high-quality, low-latency generative modeling.

We propose a generalized maximum a posteriori (MAP) channel estimation framework that exploits diffusion probabilistic priors for downlink massive multiple-input multiple-output (MIMO) systems. The channel estimation problem is reformulated as a generalized MAP inference task, where we derive the probability density of the implicit channel prior encoded by the diffusion model and obtain a simplified gradient expression of its log-prior. This leads to an efficient stochastic gradient ascent algorithm, termed Diffusion Prior Bayesian Stochastic Optimization (DP-BSO). Numerical experiments show that DP-BSO consistently outperforms classical baselines, with particularly significant gains under limited computational resources.

Robust automatic modulation recognition (AMR) under open-set conditions and adversarial attacks remains a critical challenge for secure and reliable wireless communication. This letter proposes a Boundary-aware Margin Shaping (BMS) framework that unifies open-set recognition and adversarial robustness through adaptive boundary modeling. BMS introduces a margin-aware training scheme that jointly stabilizes feature representations, calibrates confidence, and smooths decision boundaries under adversarial conditions. For inference, we employ a standard KNN detector to perform out-of-distribution (OOD) separation. Comprehensive experiments on the RML2016.10a dataset demonstrate that BMS consistently achieves superior classification and OOD detection performance across various SNRs and adversarial attacks.

Representation Autoencoders (RAEs) have improved diffusion and flow models by semantically richer latent space owing to the strongly label-wise clustered DINO features in the pretrained encoders. Yet in the distillation stage, the severe anisotropy and large curvatures caused by the rich semantic representations would hinder the convergence and performance, making the trajectory-based distillation unstable. In this work, we argue that the RAE latent space is compatible with distillation via the newly proposed Drifting Models. We first quantitatively study the curvatures and isotropy statistics across different autoencoders, and theoretically reveal that Drifting Model itself is highly likely to fail on extremely scattered spaces like reconstruction-based VAEs. These motivate us to apply the drifting paradigm directly to representation autoencoders. Our proposed method, Drift-RAE, distills pretrained flow models in RAE latent spaces using Drifting, together with insightful modifications that improve training stability by thereotically aligning drifting fields with other frameworks. Regarding the experimental evidences, we achieve 1.77 FID on ImageNet 256 dataset using only 10k distillation steps, surpassing state-of-the-art RAE distillation methods and appearing comparative with the original Drifting Model without requiring an auxiliary MAE feature extractor. The code will be made publicly available.

Most existing change detection (CD) methods focus on optical images captured at different times, and deep learning (DL) has achieved remarkable success in this domain. However, in extreme scenarios such as disaster response, synthetic aperture radar (SAR), with its active imaging capability, is more suitable for providing post-event data. This introduces new challenges for CD methods, as existing weight-sharing Siamese networks struggle to effectively learn the cross-modal data distribution between optical and SAR images. To address this challenge, we propose a unified MultiModal CD framework, M^2CD. We integrate Mixture of Experts (MoE) modules into the backbone to explicitly handle diverse modalities, thereby enhancing the model's ability to learn multimodal data distributions. Additionally, we innovatively propose an Optical-to-SAR guided path (O2SP) and implement self-distillation during training to reduce the feature space discrepancy between different modalities, further alleviating the model's learning burden. We design multiple variants of M^2CD based on both CNN and Transformer backbones. Extensive experiments validate the effectiveness of the proposed framework, with the MiT-b1 version of M^2CD outperforming all state-of-the-art (SOTA) methods in optical-SAR CD tasks.

Diffusion models excel at generating high-quality, diverse images but suffer from training data memorization, raising critical privacy and safety concerns. Data unlearning has emerged to mitigate this issue by removing the influence of specific data without retraining from scratch. We propose ReTrack, a fast and effective data unlearning method for diffusion models. ReTrack employs importance sampling to construct a more efficient fine-tuning loss, which we approximate by retaining only dominant terms. This yields an interpretable objective that redirects denoising trajectories toward the k-nearest neighbors, enabling efficient unlearning while preserving generative quality. Experiments on MNIST T-Shirt, CelebA-HQ, CIFAR-10, and Stable Diffusion show that ReTrack achieves state-of-the-art performance, striking the best trade-off between unlearning strength and generation quality preservation.

Diffusion-based inverse algorithms have shown remarkable performance across various inverse problems, yet their reliance on numerous denoising steps incurs high computational costs. While recent developments of fast diffusion ODE solvers offer effective acceleration for diffusion sampling without observations, their application in inverse problems remains limited due to the heterogeneous formulations of inverse algorithms and their prevalent use of approximations and heuristics, which often introduce significant errors that undermine the reliability of analytical solvers. In this work, we begin with an analysis of ODE solvers for inverse problems that reveals a linear combination structure of approximations for the inverse trajectory. Building on this insight, we propose a canonical form that unifies a broad class of diffusion-based inverse algorithms and facilitates the design of more generalizable solvers. Inspired by the linear subspace search strategy, we propose Learnable Linear Extrapolation (LLE), a lightweight approach that universally enhances the performance of any diffusion-based inverse algorithm conforming to our canonical form. LLE optimizes the combination coefficients to refine current predictions using previous estimates, alleviating the sensitivity of analytical solvers for inverse algorithms. Extensive experiments demonstrate consistent improvements of the proposed LLE method across multiple algorithms and tasks, indicating its potential for more efficient solutions and boosted performance of diffusion-based inverse algorithms with limited steps. Codes for reproducing our experiments are available at https://github.com/weigerzan/LLE_inverse_problem.

Momentum accelerated stochastic gradient descent (SGDM) has gained significant popularity in several signal processing and machine learning tasks. Despite its widespread success, the step size of SGDM remains a critical hyperparameter affecting its performance and often requires manual tuning. Recently, some works have introduced the Polyak step size to SGDM and provided corresponding convergence analysis. However, the convergence guarantee of existing Polyak step sizes for SGDM are limited to convex objectives and lack theoretical support for more widely applicable non-convex problems. To bridge this gap, we design a novel Polyak adaptive step size for SGDM. The proposed algorithm, termed SGDM-APS, incorporates a moving average form tailored for the momentum mechanism in SGDM. We establish the convergence guarantees of SGDM-APS for both convex and non-convex objectives, providing theoretical analysis of its effectiveness. To the best of our knowledge, SGDM-APS is the first Polyak step size for SGDM with general convergence guarantee. Our analysis can also be extended to constant step size SGDM, enriching the theoretical comprehension of the classic SGDM algorithm. Through extensive experiments on diverse benchmarks, we demonstrate that SGDM-APS achieves competitive convergence rates and generalization performance compared to several popular optimization algorithms.

Deep learning (DL) has achieved remarkable success in the field of change detection (CD) for remote sensing images. Existing training methods for DL-based CD models are predominantly single-stage, single-stream, and end-to-end. Despite the numerous optimization techniques proposed, such as advanced network architectures, loss functions, and hyperparameter tuning, these methods still struggle to achieve consistent and satisfactory detection results across images with varying change area ratios (CARs). This raises the critical question: Is the current training paradigm truly optimal? To address this question, we propose a novel multibranch self-distillation training framework. In this framework, different partition branches learn detection patterns under diverse CAR scenarios and guide the main branch through distillation. Our approach consistently enhances the detection accuracy of CD models across various change regions without introducing additional time or computational costs during the inference phase. Extensive experiments on the JL1-CD, SYSU-CD, and CDD datasets demonstrate that the MBSD framework consistently improves the performance of CD models with diverse network architectures and parameter sizes, achieving new state-of-the-art results.

Factorizing a low-rank matrix into two matrix factors with low dimensions from its noisy observations is a classical but challenging problem arising from real-world applications. This paper develops decentralized matrix factorization algorithms, i.e., factorizing a matrix whose columns are stored distributively over a network without a central agent. The performance of existing algorithms relies heavily on the accuracy of the matrix rank estimate. However, acquiring an accurate estimate is difficult in the distributed setting. In this paper and its Part II, we address this problem by introducing a novel regularization into the objective function to induce the solution with correct rank. Based on this, we propose a rank-adaptive decentralized MF algorithm. In Part I, we delineate the algorithm development from the centralized with known rank, decentralized with known rank, to the rank-adaptive decentralized settings. For the centralized algorithm, we present the first globally linear convergence analysis for the alternating gradient descent method. In the Part II, we analyze conditions for which the proposed rank-adaptive decentralized MF algorithm converges to the global solution with the correct rank. Numerical experiments based on both synthetic and real-world datasets are presented in this paper to demonstrate the effectiveness of the proposed algorithms and corroborate the theoretical claims.

Time Series Anomaly Detection (TSAD) methods often require large datasets and extensive domain knowledge, limiting their effectiveness in data-scarce environments. We introduce a novel unsupervised TSAD approach based on path signatures, which efficiently extracts shape features from time series data without requiring additional parameters. By leveraging multi-scale signature feature extraction, our method achieves faster inference and higher performance compared to deep learning models, as demonstrated on multiple cross-domain datasets. Experimental results confirm the strong generalization capability and computational efficiency of our approach.