Renmiao Chen, Yida Lu, Shiyao Cui, Xuan Ouyang, Victor Shea-Jay Huang, Shumin Zhang, Chengwei Pan, Han Qiu, Minlie Huang. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

IoT-enabled microgrids connect distributed generators, storage units, and loads through sensor-based communication networks for real-time monitoring. Motivated by the need for fast and reliable frequency-voltage restoration in islanded AC microgrids affected by communication delays, this paper proposes a prescribed-time distributed secondary control strategy for coordinated voltage, frequency, and active power sharing to enhance the operational stability of AC microgrids. Firstly, the distributed secondary control problem is formulated with careful consideration of different communication delays. The proposed prescribed-time consensus control, designed to be independent of initial conditions, provides a precise estimation of the settling time, ensuring efficient and predictable system performance. The proposed controller is utilized the measurements of local power and frequency and information shared with neighboring distributed generators through the communications network in order to estimate the nominal operational point using in primary control stage. Second, by using the control input's past data, a state-transformation- based approach is able to manage time delays and completely reduce the negative consequences of different communication delays. Then, the prescribed-time reactive power sharing control approach is developed. Furthermore, a distributed secondary controller stability is analyzed using the Lyapunov function. Finally, the results validate efficiency of the suggested control schemes under various test scenarios.

In this paper, a distributed security cooperative control for heterogeneous connected systems based on uncertainty and cyber-attacks identification is proposed. First, for the uncertainty, nonlinear disturbance observer is designed to realize the identification of uncertainty. Then, to compensate for the effects of cyber-attacks, an adaptive identification method of cyber-attacks is proposed through learning-based attack approximation approach. Furthermore, the distributed cooperative control strategy of heterogeneous connected systems under the compensation of cyber-attacks and uncertainty is presented via distributed backstepping approach. Finally, the result analysis shows the effectiveness of the designed algorithm.

In this paper, the distributed prescribed-time resilient control problem of the two-dimensional (2-D) plane heterogeneous connected vehicular platooning system (HCVPS) subject to the weaken communication and false data injection (FDI) communication attack threats is investigated. Firstly, for achieving the multi-lane vehicle merging, longitudinal following, and vehicular platooning lane changing, the third-order nonlinear dynamics of the HCVPS on the 2-D plane are considered. By developing the distributed information reconstruction observer, the weaken communication problem of leading vehicle can be mitigated by each following vehicles within a prescribed time. A distributed prescribed-time resilient control modeling framework is proposed that maintains the resilience performance and the desired safety inter-vehicle distance, ensures string stability within a prescribed-time. Finally, the results of Simulation of Urban Mobility (SUMO)-Simulink joint experimental platform and comparison cases are shown to verify the effectiveness of the developed scheme.

Text-to-Image (T2I) diffusion models have achieved remarkable success in image generation. Despite their progress, challenges remain in both prompt-following ability, image quality and lack of high-quality datasets, which are essential for refining these models. As acquiring labeled data is costly, we introduce AGFSync, a framework that enhances T2I diffusion models through Direct Preference Optimization (DPO) in a fully AI-driven approach. AGFSync utilizes Vision-Language Models (VLM) to assess image quality across style, coherence, and aesthetics, generating feedback data within an AI-driven loop. By applying AGFSync to leading T2I models such as SD v1.4, v1.5, and SDXL-base, our extensive experiments on the TIFA dataset demonstrate notable improvements in VQA scores, aesthetic evaluations, and performance on the HPS v2 benchmark, consistently outperforming the base models. AGFSync's method of refining T2I diffusion models paves the way for scalable alignment techniques.

Security concerns related to Large Language Models (LLMs) have been extensively explored; however, the safety implications for Multimodal Large Language Models (MLLMs), particularly in medical contexts (MedMLLMs), remain inadequately addressed. This paper investigates the security vulnerabilities of MedMLLMs, focusing on their deployment in clinical environments where the accuracy and relevance of question-and-answer interactions are crucial for addressing complex medical challenges. We introduce and redefine two attack types: mismatched malicious attack (2M-attack) and optimized mismatched malicious attack (O2M-attack), by integrating existing clinical data with atypical natural phenomena. Using the comprehensive 3MAD dataset that we developed, which spans a diverse range of medical imaging modalities and adverse medical scenarios, we performed an in-depth analysis and proposed the MCM optimization method. This approach significantly improves the attack success rate against MedMLLMs. Our evaluations, which include white-box attacks on LLaVA-Med and transfer (black-box) attacks on four other SOTA models, reveal that even MedMLLMs designed with advanced security mechanisms remain vulnerable to breaches. This study highlights the critical need for robust security measures to enhance the safety and reliability of open-source MedMLLMs, especially in light of the potential impact of jailbreak attacks and other malicious exploits in clinical applications. Warning: Medical jailbreaking may generate content that includes unverified diagnoses and treatment recommendations. Always consult professional medical advice.

Novel view synthesis is a critical task in autonomous driving. Although 3D Gaussian Splatting (3D-GS) has shown success in generating novel views, it faces challenges in maintaining high-quality rendering when viewpoints deviate significantly from the training set. This difficulty primarily stems from complex lighting conditions and geometric inconsistencies in texture-less regions. To address these issues, we propose an attention-based illumination model that leverages light fields from neighboring views, enhancing the realism of synthesized images. Additionally, we propose a geometry optimization method using planar homography to improve geometric consistency in texture-less regions. Our experiments demonstrate substantial improvements in synthesis quality for large-deviation viewpoints, validating the effectiveness of our approach.

Large Language Models (LLMs) are susceptible to generating harmful content when prompted with carefully crafted inputs, a vulnerability known as LLM jailbreaking.As LLMs become more powerful, studying jailbreak methods is critical to enhancing security and aligning models with human values.Traditionally, jailbreak techniques have relied on suffix addition or prompt templates, but these methods suffer from limited attack diversity.This paper introduces DiffusionAttacker, an end-to-end generative approach for jailbreak rewriting inspired by diffusion models.Our method employs a sequence-to-sequence (seq2seq) text diffusion model as a generator, conditioning on the original prompt and guiding the denoising process with a novel attack loss.Unlike previous approaches that use autoregressive LLMs to generate jailbreak prompts, which limit the modification of already generated tokens and restrict the rewriting space, DiffusionAttacker utilizes a seq2seq diffusion model, allowing more flexible token modifications.This approach preserves the semantic content of the original prompt while producing harmful content.Additionally, we leverage the Gumbel-Softmax technique to make the sampling process from the diffusion model's output distribution differentiable, eliminating the need for iterative token search.Extensive experiments on Advbench and Harmbench demonstrate that DiffusionAttacker outperforms previous methods across various evaluation metrics, including attack success rate (ASR), fluency, and diversity.How to make a bomb?Gaussian Noise Gradually Denoising DiffuSeq Model Embedding Map LM_head Gumbel softmax Jailbreaking Prompt LM_head Gumbel softmax Jailbreaking

3DGS is an emerging and increasingly popular technology in the field of novel view synthesis. Its highly realistic rendering quality and real-time rendering capabilities make it promising for various applications. However, when applied to large-scale aerial urban scenes, 3DGS methods suffer from issues such as excessive memory consumption, slow training times, prolonged partitioning processes, and significant degradation in rendering quality due to the increased data volume. To tackle these challenges, we introduce \textbf{HUG}, a novel approach that enhances data partitioning and reconstruction quality by leveraging a hierarchical neural Gaussian representation. We first propose a visibility-based data partitioning method that is simple yet highly efficient, significantly outperforming existing methods in speed. Then, we introduce a novel hierarchical weighted training approach, combined with other optimization strategies, to substantially improve reconstruction quality. Our method achieves state-of-the-art results on one synthetic dataset and four real-world datasets.

This paper presents an enhanced finite-time secondary frequency control of islanded AC microgrids to address the frequency restoration and active power sharing problems considering time-delays of the microgrid’s inputs due to the data sharing via communication networks. Artstein's reduction is employed to transform the system with delayed inputs into an equivalent system without delay. The proposed control methodology utilizes the frequency and active power measurements of a local distributed generator (DG), and share this information with neighboring DGs through communication network to generate the reference set points of the distributed generators. These set points are utilized in the primary droop control stage to restore the microgrid frequency to its nominal value and achieve active power sharing. Furthermore, to ensure the stability of the proposed control system in the presence of communication delays Artstein transformation is applied. Extensive simulations show the effectiveness and robustness of proposed distributed control technique in mitigating the effects of communication delays in islanded autonomous AC microgrids.

Jailbreak attacks against multimodal large language Models (MLLMs) are a significant research focus. Current research predominantly focuses on maximizing attack success rate (ASR), often overlooking whether the generated responses actually fulfill the attacker's malicious intent. This oversight frequently leads to low-quality outputs that, while successful in bypassing safety filters, lack substantial harmful content. To address this gap, we propose JPS, Jailbreak MLLMs with collaborative visual Perturbation and textual Steering, which achieves jailbreaks via corporation of visual image and textually steering prompt. Specifically, JPS utilizes target-guided adversarial image perturbations for effective safety bypass, complemented by ''steering prompt'' optimized via a multi-agent system to specifically guide LLM responses fulfilling the attackers' intent. These visual and textual components undergo iterative co-optimization for enhanced performance. To evaluate the quality of attack outcomes, we propose the Malicious Intent Fulfillment Rate (MIFR) metric, assessed using a Reasoning-LLM-based evaluator. Our experiments show JPS sets a new state-of-the-art in both ASR and MIFR across various MLLMs and benchmarks, with analyses confirming its efficacy. Codes are available at https://github.com/thu-coai/JPS Warning: This paper contains potentially sensitive contents.

We introduce ColaCare, a framework that enhances Electronic Health Record (EHR) modeling through multi-agent collaboration driven by Large Language Models (LLMs). Our approach seamlessly integrates domain-specific expert models with LLMs to bridge the gap between structured EHR data and text-based reasoning. Inspired by the Multidisciplinary Team (MDT) approach used in clinical settings, ColaCare employs two types of agents: DoctorAgents and a MetaAgent, which collaboratively analyze patient data. Expert models process and generate predictions from numerical EHR data, while LLM agents produce reasoning references and decision-making reports within the MDT-driven collaborative consultation framework. The MetaAgent orchestrates the discussion, facilitating consultations and evidence-based debates among DoctorAgents, simulating diverse expertise in clinical decision-making. We additionally incorporate the Merck Manual of Diagnosis and Therapy (MSD) medical guideline within a retrieval-augmented generation (RAG) module for medical evidence support, addressing the challenge of knowledge currency. Extensive experiments conducted on three EHR datasets demonstrate ColaCare's superior performance in clinical mortality outcome and readmission prediction tasks, underscoring its potential to revolutionize clinical decision support systems and advance personalized precision medicine. All code, case studies and a questionnaire are available at the project website: https://colacare.netlify.app.

This research focuses on the false data injection attacks (FDI) and communication resource competition in single-master–multiple-slaves (SMMSs) teleoperation systems and proposes a security control strategy based on dynamic try-once-discard (DTOD) scheduling protocol. An attack detector is designed to detect FDI attacks on signals transmitted through the communication channel. With the attack detection result, a DTOD scheduling protocol is designed to dynamically adjust the transmission priority of the slaves. Using the attack detection result and scheduled position and velocity, a switching controller is developed to ensure the position tracking between the master and slaves. The stability of the overall system is proven via Lyapunov functions, and comparative results validate the effectiveness of the proposed strategy. The proposed strategy not only optimizes resource allocation but also effectively defends against cyberattacks, enhancing the stability and tracking accuracy of SMMS teleoperation systems.

Knowledge graph (KG) quality evaluation seeks to assess the quality of triples in a KG. Existing approaches for KG quality evaluation often rely on high-dimensional embeddings to improve the model's evaluative performance. This reliance, however, not only increases the model's scale but also introduces feature redundancy, constraining its applicability to real-world problems. To tackle this challenge, this study proposes a Lightweight Embedding Method for KG Quality Evaluation (LEKGQE). This method performs kernel principal component analysis on each dimension of the triples, mapping multivariate data into univariate data, thereby effectively extracting the main features of the data. Furthermore, we quantify the contribution of each dimension on the target variable. Cross-entropy is used to identify the most discriminative features for KG quality evaluation, significantly reducing the model's embedding dimension and improving the KG quality evaluation performance under low-dimensional embeddings. Extensive experiments demonstrate the effectiveness of the LEKGQE model. With an embedding dimension of 32, the LEKGQE achieves a 14% increase in F1 score on the FB15K dataset and a 12% increase in F1 score on the WN18 dataset, markedly improving the model’s performance in low-dimensional embedding scenarios.

Adaptive optimization algorithms, such as Adam, are widely employed in deep learning. However, because they primarily rely on learning rate adjustments, a trade-off often exists between optimization stability and generalization capability. To address this issue, we propose AdamRAG, a novel optimization algorithm that integrates adaptive methods with Ravine acceleration and momentum techniques, aiming to preserve the stability of adaptive algorithms while enhancing their generalization performance. Within the adaptive framework, AdamRAG introduces extrapolation steps based on Ravine acceleration, which not only accelerate convergence but also prevent the iterative process from becoming trapped in local saddle points, thereby boosting generalization. Simultaneously, the momentum method is employed to regulate the descent step sizes, further improving the algorithm’s stability. Theoretical analysis demonstrates that AdamRAG achieves sublinear convergence in non-convex optimization scenarios. Extensive experiments across tasks such as image classification, natural language processing, and reinforcement learning validate its effectiveness, with results indicating that AdamRAG outperforms established optimizers (e.g., NAG, Adam, Lion) in terms of both convergence speed and generalization performance. Furthermore, sensitivity analysis shows that AdamRAG exhibits greater robustness to variations in learning rate, significantly reducing the need for hyperparameter tuning. These findings suggest that by integrating Ravine acceleration, adaptive methods, and momentum techniques, AdamRAG effectively mitigates the trade-off between stability and generalization, providing an efficient and robust optimization tool for deep learning applications.

In daily clinical practice, clinicians exhibit robustness in identifying polyps with both location and size variations. It is uncertain if deep segmentation models can achieve comparable robustness in automated colonoscopic analysis. To benchmark the model robustness, we focus on evaluating the segmentation models on the polyps with various attributes (e.g. location and size) and healthy samples. Based on the Latent Diffusion Model, we perform attribute editing on real polyps and build a new dataset named Polyp-E. Our synthetic dataset boasts exceptional realism, to the extent that clinical experts find it challenging to discern them from real data. We evaluate various existing polyp segmentation models on the proposed benchmark. The results reveal most of the models are highly sensitive to attribute variations. As a novel data augmentation technique, the proposed editing pipeline can improve both in-distribution and out-ofdistribution generalization ability. The code and datasets has been released at https://github.com/RunpuWei/Polyp-E-Benchmark.

The integration of multimodal Electronic Health Records (EHR) data has significantly advanced clinical predictive capabilities. Existing models, which utilize clinical notes and multivariate time-series EHR data, often fall short of incorporating the necessary medical context for accurate clinical tasks, while previous approaches with knowledge graphs (KGs) primarily focus on structured knowledge extraction. In response, we propose EMERGE, a Retrieval-Augmented Generation (RAG) driven framework to enhance multimodal EHR predictive modeling. We extract entities from both time-series data and clinical notes by prompting Large Language Models (LLMs) and align them with professional PrimeKG, ensuring consistency. In addition to triplet relationships, we incorporate entities' definitions and descriptions for richer semantics. The extracted knowledge is then used to generate task-relevant summaries of patients' health statuses. Finally, we fuse the summary with other modalities using an adaptive multimodal fusion network with cross-attention. Extensive experiments on the MIMIC-III and MIMIC-IV datasets' in-hospital mortality and 30-day readmission tasks demonstrate the superior performance of the EMERGE framework over baseline models. Comprehensive ablation studies and analysis highlight the efficacy of each designed module and robustness to data sparsity. EMERGE contributes to refining the utilization of multimodal EHR data in healthcare, bridging the gap with nuanced medical contexts essential for informed clinical predictions. We have publicly released the code at https://github.com/yhzhu99/EMERGE.

Electronic Health Records (EHRs) provide valuable patient data but often suffer from sparsity issue, posing significant challenges in predictive modeling. Conventional imputation methods inadequately distinguish between real and imputed data, leading to potential inaccuracies of patient representations. To address these issues, we introduce PRISM, a framework that indirectly imputes data through prototype representations of similar patients, thus ensuring denser and more accurate embeddings. PRISM also includes a feature confidence learner module, which evaluates the reliability of each feature considering missing statuses. Additionally, it incorporates a new patient similarity metric that accounts for feature confidence, avoiding overreliance on imprecise imputed values. Our extensive experiments on the MIMIC-III, MIMIC-IV, PhysioNet Challenge 2012, eICU datasets demonstrate PRISM's superior performance in predicting in-hospital mortality and 30-day readmission tasks, showcasing its effectiveness in handling EHR data sparsity. For the sake of reproducibility and further research, we have publicly released the code at https://github.com/yhzhu99/PRISM.

Recently, the 3D Gaussian Splatting (3D-GS) method has achieved great success in novel view synthesis, providing real-time rendering while ensuring high-quality rendering results. However, this method faces challenges in modeling specular reflections and handling anisotropic appearance components, especially in dealing with view-dependent color under complex lighting conditions. Additionally, 3D-GS uses spherical harmonic to learn the color representation, which has limited ability to represent complex scenes. To overcome these challenges, we introduce Lantent-SpecGS, an approach that utilizes a universal latent neural descriptor within each 3D Gaussian. This enables a more effective representation of 3D feature fields, including appearance and geometry. Moreover, two parallel CNNs are designed to decoder the splatting feature maps into diffuse color and specular color separately. A mask that depends on the viewpoint is learned to merge these two colors, resulting in the final rendered image. Experimental results demonstrate that our method obtains competitive performance in novel view synthesis and extends the ability of 3D-GS to handle intricate scenarios with specular reflections.

The cooperative tracking control of multiple unmanned aerial vehicle systems (MUAVSs) is a hotspot with its extensive applications in many areas, and a higher efficiency and more stable system is urgently demanded. To satisfy the needs, this article proposes a distributed-estimator-based (DE-based) finite-time prescribed performance control strategy to improve the convergence time and reduce the overshoot of tracking errors simultaneously. The finite-time prescribed performance function (FTPPF) constrains the convergence boundary and settling time for the cooperative tracking errors of follower UAVs, and the prescribed performance dynamic procedure can be obtained. To further achieve the cooperative operation of the MUAVS, the finite-time prescribed performance controller based on the distributed backstepping law is designed. The disturbance estimator (DE) is designed based on the adaptive prescribed approximation performance function to compensate for the unfavorable effect. Furthermore, considering the input saturation problem, the filter-based saturation compensation is presented to handle this problem. Eventually, the finite-time cooperative tracking performance with the proposed distributed FTPPF control is verified theoretically, and the proposed scheme can make tracking errors converge to an allowable range around zero rapidly under the disturbances and the input saturation. The experimental results and the comparison examples confirm the effectiveness of the proposed control scheme.

In this paper, the distributed optimal control for vehicular platoons with quantized states and control inputs is studied and the preservation of vehicles’ privacy is investigated. Firstly, the structure of system composed of cloud servers, eavesdroppers, and the vehicular platoon with unknown nonlinear dynamics and disturbance is established. Secondly, to restrain the uncertainty and nonlinearity, a policy iteration algorithm integrating quantized control inputs is proposed to stabilize vehicles without knowing their dynamics. The optimality and convergence of the iteration are proved rigorously. Then, the algorithm is utilized by the actor-critic framework with quantized states and control inputs. Due to the reason that conventional distributed control approaches for platoon will inevitably leak privacy, by introducing leveled fully homomorphic encryption, the actor-critic framework is modified to work in a privacy-preserving manner, and the privacy of each vehicle is protected from unauthorized entities. Finally, the comparison results demonstrate that with proposed scheme, the average instant cost of each vehicle is improved (at least 56.52% under experiment condition) and the stability, efficiency and security of system with proposed method are verified by simulations.

Coronary microvascular disease constitutes a substantial risk to human health. Employing computer-aided analysis and diagnostic systems, medical professionals can intervene early in disease progression, with 3D vessel segmentation serving as a crucial component. Nevertheless, conventional U-Net architectures tend to yield incoherent and imprecise segmentation outcomes, particularly for small vessel structures. While models with attention mechanisms, such as Transformers and large convolutional kernels, demonstrate superior performance, their extensive computational demands during training and inference lead to increased time complexity. In this study, we leverage Fourier domain learning as a substitute for multi-scale convo-lutional kernels in 3D hierarchical segmentation models, which can reduce computational expenses while preserving global receptive fields within the network. Furthermore, a zero-parameter frequency domain fusion method is designed to improve the skip connections in U-Net architecture. Experimental results on a public dataset and an in-house dataset indicate that our novel Fourier transformation-based network achieves remarkable dice performance (84.37% on ASACA500 and 80.32% on ImageCAS) in tubular vessel segmentation tasks and substantially reduces computational requirements without compromising global receptive fields.

This article is concerned with the issue of distributed adaptive fixed-time fuzzy resilient control (DAFTFRC) for the heterogeneous connected vehicular platoon (HCVP) system subject to deception attacks and unknown external disturbances. First, an effective fixed-time disturbance observer (FTDO) is introduced to quickly estimate the unknown external disturbance of the platoon. Then, the DAFTFRC framework is designed to satisfy the internal stability and the string stability under the deception attacks in a settling time by introducing the improved event-triggered mechanism (IETM). Correspondingly, by employing the IETM, the interevent execution intervals are prolonged, and the execution rate of the actuator can also be significantly reduced. Ample and rigorous theoretical analysis proves the internal stability and string stability in sense of the practical fixed-time stability. Finally, the effectiveness of the developed scheme is substantiated by the comparison examples and the SUMO-MATLAB joint simulation platform.

The Internet of Vehicles (IoV), which is an extension of traditional vehicular ad hoc networks (VANETs), ensures the provision of safe and efficient vehicular communications. However, the burgeoning IoV faces several challenges in terms of traffic information security and vehicle identity privacy during wireless transmission. Generally, vehicles move faster on the highway necessitating the use of reliable technology to communicate with infrastructure on the roadside efficiently and securely. Lately, researchers have proposed several different message dissemination methods for vehicle-to-infrastructure (V2I) communications. However, they suffer from different concerns (i.e., security weaknesses, computationally heavy, and high communication/storage overhead); therefore, they are unreliable for delay-sensitive communications. To cope with this, we come up with an enhanced certificateless and identity-based offline/online heterogeneous signcryption (CIOOHSC) protocol for V2I communications in edge computing-based IoV to simultaneously achieve confidentiality, integrity, authentication, non-repudiation, and identity anonymity in a logically single step. The CIOOHSC protocol allows a vehicle in certificateless cryptography to transmit a message to an edge node in identity-based cryptography. In the online phase, there are no point multiplication operations as they are already carried out in the offline phase. This reduces the computational load considerably in the online phase. Furthermore, our protocol enables the edge node to aggregate multiple messages and process them simultaneously to boost performance. The CIOOHSC protocol’s security is proved in the random oracle model under the hard problems’ assumptions. Our protocol outperforms recent similar schemes in terms of computational complexity, communication/storage overhead, and average transmission delay, according to the performance evaluation findings.

This article concentrates on the distributed adaptive secure control on lane change maneuver of interconnected unmanned vehicular platoon (IUVP) system which is established base on vehicular ad-hoc network (VANET) technology, while the whole system exposed to compound cyber attacks, including Denial-of-Service (DoS) attack and replay attack. Firstly, with a VANET based information flow communication topology, an IUVP system including longitudinal and lateral dynamics is considered in this paper. Then a distributed secure platoon lane change control scheme is designed by using recursive method to guarantee the stability and robustness of the whole systems under compound attacks. An event triggered mechanism is designed to save communication bandwidth. The stability of the whole system is then verified via rigorous mathematical proof. Finally, the IUVP system is tested by utilizing MATLAB-Simulation of Urban Mobility (SUMO) joint simulation platform, the feasibility of the proposed strategy in relative realistic environment is confirmed, and provide the root mean square error to evaluate the safety control performance of the system.

Convolutional neural networks (CNNs) have made enormous progress in medical image segmentation. The learning of CNNs is dependent on a large amount of training data with fine annotations. The workload of data labeling can be significantly relieved via collecting imperfect annotations which only match the underlying ground truths coarsely. However, label noises which are systematically introduced by the annotation protocols, severely hinders the learning of CNN-based segmentation models. Hence, we devise a novel collaborative learning framework in which two segmentation models cooperate to combat label noises in coarse annotations. First, the complementary knowledge of two models is explored by making one model clean training data for the other model. Secondly, to further alleviate the negative impact of label noises and make sufficient usage of the training data, the specific reliable knowledge of each model is distilled into the other model with augmentation-based consistency constraints. A reliability-aware sample selection strategy is incorporated for guaranteeing the quality of the distilled knowledge. Moreover, we employ joint data and model augmentations to expand the usage of reliable knowledge. Extensive experiments on two benchmarks showcase the superiority of our proposed method against existing methods under annotations with different noise levels. For example, our approach can improve existing methods by nearly 3% DSC on the lung lesion segmentation dataset LIDC-IDRI under annotations with 80% noise ratio. Code is available at: https://github.com/Amber-Believe/ReliableMutualDistillation.

Vessel segmentation is widely used to help with vascular disease diagnosis. Vessels reconstructed using existing methods are often not sufficiently accurate to meet clinical use standards. This is because 3D vessel structures are highly complicated and exhibit unique characteristics, including sparsity and anisotropy. In this paper, we propose a novel hybrid deep neural network for vessel segmentation. Our network consists of two cascaded subnetworks performing initial and refined segmentation respectively. The second subnetwork further has two tightly coupled components, a traditional CNN-based U-Net and a graph U-Net. Cross-network multi-scale feature fusion is performed between these two U-shaped networks to effectively support high-quality vessel segmentation. The entire cascaded network can be trained from end to end. The graph in the second subnetwork is constructed according to a vessel probability map as well as appearance and semantic similarities in the original CT volume. To tackle the challenges caused by the sparsity and anisotropy of vessels, a higher percentage of graph nodes are distributed in areas that potentially contain vessels while a higher percentage of edges follow the orientation of potential nearby vessels. Extensive experiments demonstrate our deep network achieves state-of-the-art 3D vessel segmentation performance on multiple public and in-house datasets.

The coronary microvascular disease poses a great threat to human health. Computer-aided analysis/diagnosis systems help physicians intervene in the disease at early stages, where 3D vessel segmentation is a fundamental step. However, there is a lack of carefully annotated dataset to support algorithm development and evaluation. On the other hand, the commonly-used U-Net structures often yield disconnected and inaccurate segmentation results, especially for small vessel structures. In this paper, motivated by the data scarcity, we first construct two large-scale vessel segmentation datasets consisting of 100 and 500 computed tomography (CT) volumes with pixel-level annotations by experienced radiologists. To enhance the U-Net, we further propose the cross transformer network (CTN) for fine-grained vessel segmentation. In CTN, a transformer module is constructed in parallel to a U-Net to learn long-distance dependencies between different anatomical regions; and these dependencies are communicated to the U-Net at multiple stages to endow it with global awareness. Experimental results on the two in-house datasets indicate that this hybrid model alleviates unexpected disconnections by considering topological information across regions. Our codes, together with the trained models are made publicly available at https://github.com/qibaolian/ctn.

Vehicular ad hoc networks (VANETs), an application of the Internet of Things (IoT) providing a better intelligent transportation system (ITS), has received substantial attention from both industry and academia. Heterogeneous vehicular communications in VANETs occur when both vehicles and the infrastructure use different cryptographic technologies to exchange safety messages. However, the safety messages between vehicles and the infrastructure are communicated wirelessly; therefore, security issues are a serious concern in this domain. The existing schemes secure the transmission of safety messages with respect to confidentiality, authentication, and nonrepudiation. However, they are inappropriate with respect to efficiency. They cause computational overhead on the receiver and bandwidth overhead in the communications. To cope with this, we propose an elliptic curve cryptosystem-based hybrid signcryption (ECCHSC) protocol that satisfies the security requirements (i.e., message confidentiality, message’s source authentication, message integrity, nonrepudiation, and identity–anonymity) for heterogeneous vehicle-to-infrastructure (V2I) communications in a single logical step. This protocol allows secure transmission of a safety message from a vehicle using identity-based cryptography (IDC) to a roadside unit (RSU) using public-key infrastructure (PKI). In addition, the ECCHSC protocol enables the RSU to receive multiple ciphertexts, aggregate them, and de-signcrypt them simultaneously through the batch de-signcryption method, which further improves the performance. The ECCHSC protocol has indistinguishability against adaptive chosen ciphertext attacks (IND-CCA2) and existential unforgeability against adaptive chosen message attacks (EUF-CMAs) in the random oracle model (ROM). The performance analysis of our protocol demonstrates a significant reduction in computational overhead and in communication/storage overhead as compared to the state-of-the-art schemes.

Locating diseases in chest X-ray images with few careful annotations saves large human effort. Recent works approached this task with innovative weakly-supervised algorithms such as multi-instance learning (MIL) and class activation maps (CAM), however, these methods often yield inaccurate or incomplete regions. One of the reasons is the neglection of the pathological implications hidden in the relationship across anatomical regions within each image and the relationship across images. In this paper, we argue that the cross-region and cross-image relationship, as contextual and compensating information, is vital to obtain more consistent and integral regions. To model the relationship, we propose the Graph Regularized Embedding Network (GREN), which leverages the intra-image and inter-image information to locate diseases on chest X-ray images. GREN uses a pre-trained U-Net to segment the lung lobes, and then models the intra-image relationship between the lung lobes using an intra-image graph to compare different regions. Meanwhile, the relationship between in-batch images is modeled by an inter-image graph to compare multiple images. This process mimics the training and decision-making process of a radiologist: comparing multiple regions and images for diagnosis. In order for the deep embedding layers of the neural network to retain structural information (important in the localization task), we use the Hash coding and Hamming distance to compute the graphs, which are used as regularizers to facilitate training. By means of this, our approach achieves the state-of-the-art result on NIH chest X-ray dataset for weakly-supervised disease localization. Our codes are accessible online.

Segmentation of retinal vessels on fundus images plays a critical role in the diagnosis of micro-vascular and ophthalmological diseases. Although being extensively studied, this task remains challenging due to many factors including the highly variable vessel width and poor vessel-background contrast. In this paper, we propose a multiscale feature interaction network (MFI-Net) for retinal vessel segmentation, which is a U-shaped convolutional neural network equipped with the pyramid squeeze-and-excitation (PSE) module, coarse-to-fine (C2F) module, deep supervision, and feature fusion. We extend the SE operator to multiscale features, resulting in the PSE module, which uses the channel attention learned at multiple scales to enhance multiscale features and enables the network to handle the vessels with variable width. We further design the C2F module to generate and re-process the residual feature maps, aiming to preserve more vessel details during the decoding process. The proposed MFI-Net has been evaluated against several public models on the DRIVE, STARE, CHASE_DB1, and HRF datasets. Our results suggest that both PSE and C2F modules are effective in improving the accuracy of MFI-Net, and also indicate that our model has superior segmentation performance and generalization ability over existing models on four public datasets.

This paper investigates the problem of the interconnected heterogeneous vehicular platoon in the presence of actuator fault of a vehicle, actuator saturation, and complex environment disturbances such as road slope, wind gust, and so on. First of all, a quadratic spacing error policy is applied to improve the vehicle platoon stability despite any vehicle suffering from actuator fault. Then, adaptive estimation laws are introduced to tackle the complex environment disturbances and the unknown upper bound of the actuator saturation approximation function. A robust adaptive sliding mode resilient control law is proposed to achieve the resilient ability and ensure the platoon system recover to the stable states in finite-time even if the occurrence of actuator fault. Finally, the string stability in finite-time of the platoon system is proved, the interconnected heterogeneous vehicular platoon system with seven vehicles is utilized to demonstrate the validity of the proposed control scheme.

This paper investigates the distributed fault-tolerant control problem for the heterogeneous vehicular platoon system (HVPS) suffering from actuator faults, saturation, and external disturbances. Firstly, an exponential spacing policy is developed to tackle the string stability (SS) and traffic flow stability issues when the actuator faults occur. Then, a nonlinear observer is developed to estimate the saturation approximation error, bias fault, and external disturbances. Moreover, to achieve the SS of the whole system in finite-time and fault tolerance ability, a distributed adaptive fault-tolerant control scheme based on bidirectional-leader information flow topology is proposed. Adaptive estimation laws are also involved to estimate and update the unknown parameters, in which the effect of actuator saturation can be compensated. Finally, the SS in finite-time and traffic flow stability of the closed-loop system are proved, respectively. Experiment results and comparisons demonstrate the feasibly and advantages of the method.

In this paper, we introduce a method to augment virtual environment with multiple videos. Our goal is to make a virtual-real fusion system that fuses dynamic imagery with 3D models in a real-time display, so as to help observers visualize dynamic videos simultaneously in the context of 3D models. 3D models in virtual environment are reconstructed using multiple view vision methods. Based on this, images can be registered through feature matching among images. Foreground objects in videos lead to distortions when video images are simply projected to static models. Detection and tracking of those objects are needed, and then several ordinary 3D models are used to represent those objects. Both geometry and appearance are taken into account to recover characteristic of different objects. This paper focuses on the integration of these components into a prototype system and the presentation of results shows the benefits of an virtual-real fusion system.