This position paper argues that post-deployment monitoring in clinical AI is underdeveloped and proposes statistically valid and label-efficient testing frameworks as a principled foundation for ensuring reliability and safety in real-world deployment. A recent review found that only 9\% of FDA-registered AI-based healthcare tools include a post-deployment surveillance plan. Existing monitoring approaches are often manual, sporadic, and reactive, making them ill-suited for the dynamic environments in which clinical models operate. We contend that post-deployment monitoring should be grounded in label-efficient and statistically valid testing frameworks, offering a principled alternative to current practices. We use the term "statistically valid" to refer to methods that provide explicit guarantees on error rates (e.g., Type I/II error), enable formal inference under pre-defined assumptions, and support reproducibility—features that align with regulatory requirements. Specifically, we propose that the detection of changes in the data and model performance degradation should be framed as distinct statistical hypothesis testing problems. Grounding monitoring in statistical rigor ensures a reproducible and scientifically sound basis for maintaining the reliability of clinical AI systems. Importantly, it also opens new research directions for the technical community---spanning theory, methods, and tools for statistically principled detection, attribution, and mitigation of post-deployment model failures in real-world settings.

As a novel communication paradigm, semantic communication (SemCom) can greatly improve communication efficiency, which has aroused extensive research by scholars worldwide. As one of the important aspects of digital communication nowadays, how to combine channel estimation with SemCom is an important research direction. In this article, based on orthogonal frequency-division multiplexing (OFDM) communication architecture, the semantic importance-aware reordering-enhanced SemCom system (SIARE-SC) is proposed, which utilizes the inequality of semantic symbols combined with channel estimation in OFDM systems to reduce the distortion caused by channel estimation interpolation error (CEIE) and further improve the signal recovery quality. To enhance the generalizability of the system, we extend the verification of the effectiveness of SIARE-SC in various scenarios with different sources, channels, and pilot patterns. Furthermore, the importance reordering method proposed in the SIARE-SC has good applicability and effectiveness, which can be used to be compatible with other SemCom systems and has a significant suppression effect on the peak-to-average power ratio (PAPR). Meanwhile, CEIE has been considered for the first time to be included in the analysis of SemCom distortion, and mathematically derive the performance expressions of SIARE-SC under different channel and pilot pattern scenarios from three perspectives, namely, channel bandwidth ratio (CBR), signal-to-noise ratio (SNR), and CEIE, to obtain the corresponding bound of performance. The proposed SIARE-SC is shown to significantly improve semantic performance in various scenarios by conducting a large number of experimental tests.

With the rapid development of Internet of Things (IoT) technology, an increasing number of resource-constrained devices operate in dynamic and heterogeneous network environments, posing challenges for efficient image transmission. Multi-user semantic communication (SC) enables reduced bandwidth consumption and enhanced noise resilience by understanding the intrinsic meaning of information and sharing common semantic features across devices, offering great potential for widespread applications in various IoT scenarios. However, current multi-users SC approaches for image transmission lack adaptability and fail to consider both content and style features, leading to degraded image reconstruction quality. Moreover, semantic redundancy among devices remains underutilized, limiting bandwidth efficiency in IoT networks. To address these limitations, in this paper, a novel multi-user content-style adaptive semantic communication system for image transmission in IoT scenarios is proposed. Specifically, a dual-branch semantic information extraction and adaptive recovery scheme is first established, which simultaneously captures and adaptively fuses semantic content and style features to improve reconstruction quality. Secondly, an adaptive common information extraction and enhanced coding module is introduced for resource-limited IoT devices, which dynamically adjusts the transmission rate based on varying channel conditions and the computational capabilities of different users, further optimizing communication performance. Finally, experimental results show that the proposed method improves peak signal-to-noise (PSNR) by at least 10% under poor SNR conditions for multi-users semantic communication, compared to baseline methods.

In this paper, we introduce an end-to-end non-orthogonal multiple access (NOMA) framework for the downlink transmission of correlated information sources in the multi-user scenario, in which the data required or transmitted by multiple users share similar content. To enhance the end-to-end transmission performance, we resort to the semantic communication paradigm and build our system based on the deep joint source-channel coding (D-JSCC) scheme. Inspired by Wyner’s common information, an information theoretical concept, the common information (CI) extraction is proposed to capture the correlation between multiple users effectively. By relaxing the constraint of the object function, equivalency can be established between common information extraction and mutual information maximization. Thereby, the Jenson-Shannon divergence (JSD) is adopted in the loss function for learning the common information representation (CIR). In order to categorize the theoretical performance limit of the proposed system, semantic synonymous mapping (SSM) based information theory is applied for analyzing the effect of correlation level and different decoding schemes on the achievable channel capacity. Specifically, the analytical expression of channel capacity under additive white Gaussian noise (AWGN) and Rayleigh channel is derived and verified by Monte-Carlo experiments. By conducting simulations on three different image datasets, it is verified that our proposed scheme can outperform a series of other state-of-the-art (SoTA) multiple access or distributed source coding (DSC) schemes under up to seven user scenarios. Besides, the visualization and ablation study results validate the effectiveness of the common information extraction.

Semantic Communication (SemCom) has opened a new paradigm in the 6G system. However, the performance of SemCom can be severely affected by time-varying path loss, channel noises, and other interference in wireless channels. Therefore, we propose a novel Semantic Prior aided Channel-adaptive Equalizing and De-noising SemCom (SP-EDNSC) framework, where adaptive elimination channel impact is regarded as an inverse problem. This inverse problem is addressed through semantic priors learned from score-based generative models cached in knowledge base. To reduce distortion while enhancing perceptual quality, we further combine autoencoders, adversarial learning and diffusion models to develop a latent diffusion-based (SP-Latent-Diff EDNSC) system within the SP-EDNSC framework. In the semantic space, the joint semantic equalizer and de-noiser module utilizes the proposed latent diffusion posterior sampling method. This method iteratively executes a modified reverse stochastic differential equation to sample clean semantic features, using the time-dependent score function of likelihood and semantic priors. The semantic priors are derived from pre-trained latent diffusion models, while the likelihood is approximated by a multivariate normal distribution. Simulations demonstrate that our scheme achieves superior performance in both distortion metrics like PSNR and SSIM, as well as in perceptual performance (LPIPS).

In communication systems, the challenge of ensuring accurate data transmission across noisy channels remains paramount. While semantic communication shows potential in improving image transmission and reconstruction, existing methods still suffer from perceptual quality degradation in high-noise environments. To address these issues, we introduce DiffSC, a novel semantic communication framework that integrates the Diffusion Probabilistic Model (DPM). Within DPM, Gaussian noise modeling is leveraged to facilitate enhanced image generation. DiffSC is trained to recover semantic information compromised during transmission, amplifying its capabilities in both image reconstruction and denoising. Additionally, we propose a Multi-Dimensional Feature Extraction Module (MFM) that employs multiple convolution kernels along with channel and spatial attention mechanisms to enrich encoding and decoding. Experimental results demonstrate that DiffSC significantly outperforms existing systems, improving SSIM by more than 19% and PSNR by more than 11% compared to Deep JSCC. Further ablation studies demonstrate the effectiveness of our proposed denoiser and feature extraction modules.

Gaussian graphical models have emerged as a powerful tool for modeling and understanding multivariate data across various domains. In this paper, we consider the problem of change localization in the Gaussian graphical model, where it is known that a change has occurred in the underlying graph structure, and the goal is to localize the change rapidly. This paradigm occurs in various applications, from cyber-physical systems and biological networks to social networks and epidemiology. We introduce a novel algorithm, dubbed FOLk-DGM (Fast Online Localization in Dynamic Graphical Models), that is both computationally efficient and performs change localization with provably low latency (time elapsed before the change is localized). We present the theoretical properties of the algorithm and complement our theoretical results with experimental results.

Two-sample tests evaluate whether two samples are realizations of the same distribution (the null hypothesis) or two different distributions (the alternative hypothesis). We consider a new setting for this problem where sample features are easily measured whereas sample labels are unknown and costly to obtain. Accordingly, we devise a three-stage framework in service of performing an effective two-sample test with only a small number of sample label queries: first, a classifier is trained with samples uniformly labeled to model the posterior probabilities of the labels; second, a novel query scheme dubbed bimodal query is used to query labels of samples from both classes, and last, the classical Friedman-Rafsky (FR) two-sample test is performed on the queried samples. Theoretical analysis and extensive experiments performed on several datasets demonstrate that the proposed test controls the Type I error and has decreased Type II error relative to uniform querying and certainty-based querying. Source code for our algorithms and experimental results is available at https://github.com/wayne0908/Label-Efficient-Two-Sample.

Visible light communication on the basis of IM/DD system has attracted enormous interest in recent years. One of the major topics to be investigated in this field is orthogonal frequency division multiplexing (OFDM). This paper proposed a novel iterative receiver for PAM-DMT based hybrid OFDM in order to enhance its performance. The concept of OFDM models and structure of conventional receiver are introduced firstly. Then the proposed iterative receiver and its computational complexity are presented. Simulation showed that under the same bit error rate (BER) of 10−4, the required signal to noise ratio (SNR) for transmitting has been reduced for about 2.5 dB. In conclusion, the proposed iterative receiver could achieve a considerable performance gain under a variety of simulation conditions, which demonstrated its potential for being applied in the visual light communication system.

Regularization is an effective way to promote the generalization performance of machine learning models. In this paper, we focus on label smoothing, a form of output distribution regularization that prevents overfitting of a neural network by softening the ground-truth labels in the training data in an attempt to penalize overconfident outputs. Existing approaches typically use cross-validation to impose this smoothing, which is uniform across all training data. In this paper, we show that such label smoothing imposes a quantifiable bias in the Bayes error rate of the training data, with regions of the feature space with high overlap and low marginal likelihood having a lower bias and regions of low overlap and high marginal likelihood having a higher bias. These theoretical results motivate a simple objective function for data-dependent smoothing to mitigate the potential negative consequences of the operation while maintaining its desirable properties as a regularizer. We call this approach Structural Label Smoothing (SLS). We implement SLS and empirically validate on synthetic, Higgs, SVHN, CIFAR-10, and CIFAR-100 datasets. The results confirm our theoretical insights and demonstrate the effectiveness of the proposed method in comparison to traditional label smoothing.

Accurately and efficiently characterizing the decision boundary of classifiers is important for problems related to model selection and meta-learning. Inspired by topological data analysis, the characterization of decision boundaries using their homology has recently emerged as a general and powerful tool. In this paper, we propose an active learning algorithm to recover the homology of decision boundaries. Our algorithm sequentially and adaptively selects which samples it requires the labels of. We theoretically analyze the proposed framework and show that the query complexity of our active learning algorithm depends naturally on the intrinsic complexity of the underlying manifold. We demonstrate the effectiveness of our framework in selecting best-performing machine learning models for datasets just using their respective homological summaries. Experiments on several standard datasets show the sample complexity improvement in recovering the homology and demonstrate the practical utility of the framework for model selection.

We present a novel computational model for simultaneous image co-saliency detection and co-segmentation that concurrently explores the concepts of saliency and objectness in multiple images. It has been shown that the co-saliency detection via aggregating multiple saliency proposals by diverse visual cues can better highlight the salient objects; however, the optimal proposals are typically region-dependent and the fusion process often leads to blurred results. Co-segmentation can help preserve object boundaries, but it may suffer from complex scenes. To address these issues, we develop a unified method that addresses co-saliency detection and co-segmentation jointly via solving an energy minimization problem over a graph. Our method iteratively carries out the region-wise adaptive saliency map fusion and object segmentation to transfer useful information between the two complementary tasks. Through the optimization iterations, sharp saliency maps are gradually obtained to recover entire salient objects by referring to object segmentation, while these segmentations are progressively improved owing to the better saliency prior. We evaluate our method on four public benchmark data sets while comparing it to the state-of-the-art methods. Extensive experiments demonstrate that our method can provide consistently higher-quality results on both co-saliency detection and co-segmentation.

Inhomogeneous color distribution and intensity impose major difficulty in fully automated histopathological image (histo-image) segmentation. In this paper, we propose a novel deep learning framework for histo-image segmentation. We innovate a noise-tolerant layer to the output layer of a deep learning image segmentation framework U-Net, which alleviates the requirement of accurately segmented training images and enables “unsupervised” histo-image segmentation by taking noisy segmentation results of traditional image segmentation algorithms as the training outputs. We implement noise-tolerant U-Net for histo-image segmentation to study Duchenne Muscular Dystrophy (DMD), a muscle degenerative disease. Performance comparison with traditional algorithms and the original U-Net demonstrates the great potential of the proposed noise-tolerant U-Net for histo-image segmentation.

In this paper, we propose a novel learning framework for the problem of domain transfer learning. We map the data of two domains to one single common space, and learn a classifier in this common space. Then we adapt the common classifier to the two domains by adding two adaptive functions to it respectively. In the common space, the target domain data points are weighted and matched to the target domain in term of distributions. The weighting terms of source domain data points and the target domain classification responses are also regularized by the local reconstruction coefficients. The novel transfer learning framework is evaluated over some benchmark cross-domain data sets, and it outperforms the existing state-of-the-art transfer learning methods.