Rui Zhang

Shuowen Zhang, Yong Zeng, Rui Zhang

Integrating the unmanned aerial vehicles (UAVs) into the cellular network is envisioned to be a promising technology to significantly enhance the communication performance of both UAVs and existing terrestrial users. In this paper, we first provide an overview on the two main paradigms in cellular UAV communications, i.e., cellular-enabled UAV communication with UAVs as new aerial users served by the ground base stations (GBSs), and UAV-assisted cellular communication with UAVs as new aerial communication platforms serving the terrestrial users. Then, we focus on the former paradigm and study a new UAV trajectory design problem subject to practical communication connectivity constraints with the GBSs. Specifically, we consider a cellular-connected UAV in the mission of flying from an initial location to a final location, during which it needs to maintain reliable communication with the cellular network by associating with one GBS at each time instant. We aim to minimize the UAV's mission completion time by optimizing its trajectory, subject to a quality-of-connectivity constraint of the GBS-UAV link specified by a minimum receive signal-to-noise ratio target. To tackle this challenging non-convex problem, we first propose a graph connectivity based method to verify its feasibility. Next, by examining the GBS-UAV association sequence over time, we obtain useful structural results on the optimal UAV trajectory, based on which two efficient methods are proposed to find high-quality approximate trajectory solutions by leveraging graph theory and convex optimization techniques. The proposed methods are analytically shown to be capable of achieving a flexible trade-off between complexity and performance, and yielding a solution that is arbitrarily close to the optimal solution in polynomial time. Finally, we make concluding remarks and point out some promising directions for future work.

Rui Zhang, Ziqi Wang, Yang Li et al.

Few-shot learning (FSL) is one of the significant and hard problems in the field of image classification. However, in contrast to the rapid development of the visible light dataset, the progress in SAR target image classification is much slower. The lack of unified benchmark is a key reason for this phenomenon, which may be severely overlooked by the current literature. The researchers of SAR target image classification always report their new results on their own datasets and experimental setup. It leads to inefficiency in result comparison and impedes the further progress of this area. Motivated by this observation, we propose a novel few-shot SAR image classification benchmark (FewSAR) to address this issue. FewSAR consists of an open-source Python code library of 15 classic methods in three categories for few-shot SAR image classification. It provides an accessible and customizable testbed for different few-shot SAR image classification task. To further understanding the performance of different few-shot methods, we establish evaluation protocols and conduct extensive experiments within the benchmark. By analyzing the quantitative results and runtime under the same setting, we observe that the accuracy of metric learning methods can achieve the best results. Meta-learning methods and fine-tuning methods perform poorly on few-shot SAR images, which is primarily due to the bias of existing datasets. We believe that FewSAR will open up a new avenue for future research and development, on real-world challenges at the intersection of SAR image classification and few-shot deep learning. We will provide our code for the proposed FewSAR at https://github.com/solarlee/FewSAR.

Mohammad R. Vedady Moghadam, Rui Zhang

In multiple-input single-output (MISO) wireless power transfer (WPT) via magnetic resonant coupling (MRC), multiple transmitters are deployed to enhance the efficiency of power transfer to the electric load at a single receiver by jointly optimizing their source currents/voltages to constructively combine the induced magnetic fields at the receiver, known as magnetic beamforming. In practice, since the transmitters (power chargers) are usually at fixed locations and the receiver (e.g. mobile phone) is desired to be freely located in a target region for wireless charging, its received power can fluctuate significantly over locations even with adaptive magnetic beamforming applied. To achieve uniform coverage, the transmitters need to be optimally placed in the region, which motivates this paper. First, we derive the optimal magnetic beamforming solution in closed-form for a distributed MISO WPT system with given locations of the transmitters and receiver to maximize the deliverable power to the receiver load subject to a given sum-power constraint at all transmitters. With the optimal magnetic beamforming solution, we then jointly optimize the locations of all transmitters to maximize the minimum power deliverable to the receiver when it is being moved over a given one-dimensional (1D) region, i.e., a line of finite length. Although the formulated node placement problem is non-convex, we propose an iterative algorithm for solving it efficiently. Extensive simulation results are provided which show the significant performance gains by the proposed design with optimized transmitter locations and magnetic beamforming as compared to other benchmark schemes with non-adaptive or heuristic currents allocation and transmitters placement. Last, we extend the node placement problem to the more general case of two-dimensional (2D) region, and draw the key insights.

Mohammad R. Vedady Moghadam, Rui Zhang

Magnetic resonant coupling (MRC) is a practically appealing method for realizing the near-field wireless power transfer (WPT). The MRC-WPT system with a single pair of transmitter and receiver has been extensively studied in the literature, while there is limited work on the general setup with multiple transmitters and/or receivers. In this paper, we consider a point-to-multipoint MRC-WPT system with one transmitter sending power wirelessly to a set of distributed receivers simultaneously. We derive the power delivered to the load of each receiver in closed-form expression, and reveal a "near-far" fairness issue in multiuser power transmission due to users' distance-dependent mutual inductances with the transmitter. We also show that by designing the receivers' load resistances, the near-far issue can be optimally solved. Specifically, we propose a centralized algorithm to jointly optimize the load resistances to minimize the power drawn from the energy source at the transmitter under given power requirements for the loads. We also devise a distributed algorithm for the receivers to adjust their load resistances iteratively, for ease of practical implementation.

Wenbo Jiang, Hongwei Li, Jiaming He et al.

Recently, deep learning-based Image-to-Image (I2I) networks have become the predominant choice for I2I tasks such as image super-resolution and denoising. Despite their remarkable performance, the backdoor vulnerability of I2I networks has not been explored. To fill this research gap, we conduct a comprehensive investigation on the susceptibility of I2I networks to backdoor attacks. Specifically, we propose a novel backdoor attack technique, where the compromised I2I network behaves normally on clean input images, yet outputs a predefined image of the adversary for malicious input images containing the trigger. To achieve this I2I backdoor attack, we propose a targeted universal adversarial perturbation (UAP) generation algorithm for I2I networks, where the generated UAP is used as the backdoor trigger. Additionally, in the backdoor training process that contains the main task and the backdoor task, multi-task learning (MTL) with dynamic weighting methods is employed to accelerate convergence rates. In addition to attacking I2I tasks, we extend our I2I backdoor to attack downstream tasks, including image classification and object detection. Extensive experiments demonstrate the effectiveness of the I2I backdoor on state-of-the-art I2I network architectures, as well as the robustness against different mainstream backdoor defenses.

Pucheng Dang, Xing Hu, Dong Li et al.

Current text-to-image (T2I) synthesis diffusion models raise misuse concerns, particularly in creating prohibited or not-safe-for-work (NSFW) images. To address this, various safety mechanisms and red teaming attack methods are proposed to enhance or expose the T2I model's capability to generate unsuitable content. However, many red teaming attack methods assume knowledge of the text encoders, limiting their practical usage. In this work, we rethink the case of \textit{purely black-box} attacks without prior knowledge of the T2l model. To overcome the unavailability of gradients and the inability to optimize attacks within a discrete prompt space, we propose DiffZOO which applies Zeroth Order Optimization to procure gradient approximations and harnesses both C-PRV and D-PRV to enhance attack prompts within the discrete prompt domain. We evaluated our method across multiple safety mechanisms of the T2I diffusion model and online servers. Experiments on multiple state-of-the-art safety mechanisms show that DiffZOO attains an 8.5% higher average attack success rate than previous works, hence its promise as a practical red teaming tool for T2l models.

Linke Ouyang, Yuan Qu, Hongbin Zhou et al.

Document content extraction is a critical task in computer vision, underpinning the data needs of large language models (LLMs) and retrieval-augmented generation (RAG) systems. Despite recent progress, current document parsing methods have not been fairly and comprehensively evaluated due to the narrow coverage of document types and the simplified, unrealistic evaluation procedures in existing benchmarks. To address these gaps, we introduce OmniDocBench, a novel benchmark featuring high-quality annotations across nine document sources, including academic papers, textbooks, and more challenging cases such as handwritten notes and densely typeset newspapers. OmniDocBench supports flexible, multi-level evaluations--ranging from an end-to-end assessment to the task-specific and attribute--based analysis using 19 layout categories and 15 attribute labels. We conduct a thorough evaluation of both pipeline-based methods and end-to-end vision-language models, revealing their strengths and weaknesses across different document types. OmniDocBench sets a new standard for the fair, diverse, and fine-grained evaluation in document parsing. Dataset and code are available at https://github.com/opendatalab/OmniDocBench.

Ryo Kamoi, Yusen Zhang, Sarkar Snigdha Sarathi Das et al.

Large Vision Language Models (LVLMs) have achieved remarkable performance in various vision-language tasks. However, it is still unclear how accurately LVLMs can perceive visual information in images. In particular, the capability of LVLMs to perceive geometric information, such as shape, angle, and size, remains insufficiently analyzed, although the perception of these properties is crucial for tasks that require a detailed visual understanding. In this work, we introduce VisOnlyQA, a dataset for evaluating the geometric perception of LVLMs, and reveal that LVLMs often cannot accurately perceive basic geometric information in images, while human performance is nearly perfect. VisOnlyQA consists of 12 tasks that directly ask about geometric information in geometric shapes, charts, chemical structures, and 3D shapes. Our experiments highlight the following findings: (i) State-of-the-art LVLMs struggle with basic geometric perception. 23 LVLMs we evaluate, including GPT-4o and Gemini 2.5 Pro, work poorly on VisOnlyQA. (ii) Additional training data does not resolve this issue. Fine-tuning on the training set of VisOnlyQA is not always effective, even for in-distribution tasks. (iii) LLM may be the bottleneck. LVLMs using stronger LLMs exhibit better geometric perception on VisOnlyQA, while it does not require complex reasoning, suggesting that the way LVLMs process information from visual encoders is a bottleneck. The datasets, code, and model responses are provided at https://github.com/psunlpgroup/VisOnlyQA.

Yichen Li, Qiyu Qin, Gaoyang Zhu et al.

Sequential recommendation is an advanced recommendation technique that utilizes the sequence of user behaviors to generate personalized suggestions by modeling the temporal dependencies and patterns in user preferences. However, it requires a server to centrally collect users' data, which poses a threat to the data privacy of different users. In recent years, federated learning has emerged as a distributed architecture that allows participants to train a global model while keeping their private data locally. This survey pioneers Federated Sequential Recommendation (FedSR), where each user joins as a participant in federated training to achieve a recommendation service that balances data privacy and model performance. We begin with an introduction to the background and unique challenges of FedSR. Then, we review existing solutions from two levels, each of which includes two specific techniques. Additionally, we discuss the critical challenges and future research directions in FedSR.

Zelin Zhu, Kai Yang, Rui Zhang

Wireless localization and sensing technologies are essential in modern wireless networks, supporting applications in smart cities, the Internet of Things (IoT), and autonomous systems. High-performance localization and sensing systems are critical for both network efficiency and emerging intelligent applications. Integrating channel state information (CSI) with deep learning has recently emerged as a promising solution. Recent works have leveraged the spatial diversity of multiple input multiple output (MIMO) systems and the frequency granularity of orthogonal frequency division multiplexing (OFDM) waveforms to improve spatial resolution. Nevertheless, the joint modeling of localization and sensing under the high-dimensional CSI characteristics of MIMO-OFDM systems remains insufficiently investigated. This work aims to jointly model and optimize localization and sensing tasks to harness their potential synergy. We first formulate localization and sensing as a mixed-integer bilevel deep learning problem and then propose a novel stochastic proximal gradient-based mixed-integer bilevel optimization (SPG-MIBO) algorithm. SPG-MIBO is well-suited for high-dimensional and large-scale datasets, leveraging mini-batch training at each step for computational and memory efficiency. The algorithm is also supported by theoretical convergence guarantees. Extensive experiments on multiple datasets validate its effectiveness and highlight the performance gains from joint localization and sensing optimization.

Ruiqi Zhang, Simon H. Tindemans

Multilevel Monte Carlo (MLMC) is a flexible and effective variance reduction technique for accelerating reliability assessments of complex power system. Recently, data-driven surrogate models have been proposed as lower-level models in the MLMC framework due to their high correlation and negligible execution time once trained. However, in resource adequacy assessments, pre-labeled datasets are typically unavailable. For large-scale systems, the efficiency gains from surrogate models are often offset by the substantial time required for labeling training data. Therefore, this paper introduces a speed metric that accounts for training time in evaluating MLMC efficiency. Considering the total time budget is limited, a vote-by-committee active learning approach is proposed to reduce the required labeling calls. A case study demonstrates that, within a given computational budget, active learning in combination with MLMC can result in a substantial reduction variance.

Yongkang Xiao, Sinian Zhang, Yi Dai et al.

Knowledge graph completion (KGC) aims to predict missing triples in knowledge graphs (KGs) by leveraging existing triples and textual information. Recently, generative large language models (LLMs) have been increasingly employed for graph tasks. However, current approaches typically encode graph context in textual form, which fails to fully exploit the potential of LLMs for perceiving and reasoning about graph structures. To address this limitation, we propose DrKGC (Dynamic Subgraph Retrieval-Augmented LLMs for Knowledge Graph Completion). DrKGC employs a flexible lightweight model training strategy to learn structural embeddings and logical rules within the KG. It then leverages a novel bottom-up graph retrieval method to extract a subgraph for each query guided by the learned rules. Finally, a graph convolutional network (GCN) adapter uses the retrieved subgraph to enhance the structural embeddings, which are then integrated into the prompt for effective LLM fine-tuning. Experimental results on two general domain benchmark datasets and two biomedical datasets demonstrate the superior performance of DrKGC. Furthermore, a realistic case study in the biomedical domain highlights its interpretability and practical utility.

Zihan Wang, Hongwei Li, Rui Zhang et al.

In this paper, we present a new form of backdoor attack against Large Language Models (LLMs): lingual-backdoor attacks. The key novelty of lingual-backdoor attacks is that the language itself serves as the trigger to hijack the infected LLMs to generate inflammatory speech. They enable the precise targeting of a specific language-speaking group, exacerbating racial discrimination by malicious entities. We first implement a baseline lingual-backdoor attack, which is carried out by poisoning a set of training data for specific downstream tasks through translation into the trigger language. However, this baseline attack suffers from poor task generalization and is impractical in real-world settings. To address this challenge, we design BadLingual, a novel task-agnostic lingual-backdoor, capable of triggering any downstream tasks within the chat LLMs, regardless of the specific questions of these tasks. We design a new approach using PPL-constrained Greedy Coordinate Gradient-based Search (PGCG) based adversarial training to expand the decision boundary of lingual-backdoor, thereby enhancing the generalization ability of lingual-backdoor across various tasks. We perform extensive experiments to validate the effectiveness of our proposed attacks. Specifically, the baseline attack achieves an ASR of over 90% on the specified tasks. However, its ASR reaches only 37.61% across six tasks in the task-agnostic scenario. In contrast, BadLingual brings up to 37.35% improvement over the baseline. Our study sheds light on a new perspective of vulnerabilities in LLMs with multilingual capabilities and is expected to promote future research on the potential defenses to enhance the LLMs' robustness

Jie Peng, Shuang Zhou, Longwei Yang et al.

Cancer prognosis is a critical task that involves predicting patient outcomes and survival rates. To enhance prediction accuracy, previous studies have integrated diverse data modalities, such as clinical notes, medical images, and genomic data, leveraging their complementary information. However, existing approaches face two major limitations. First, they struggle to incorporate newly arrived data with varying distributions into training, such as patient records from different hospitals, thus rendering sub-optimal generalizability and limited utility in real-world applications. Second, most multimodal integration methods rely on simplistic concatenation or task-specific pipelines, which fail to capture the complex interdependencies across modalities. To address these, we propose a continually evolving multi-modal foundation model. Extensive experiments on the TCGA dataset demonstrate the effectiveness of our approach, highlighting its potential to advance cancer prognosis by enabling robust and adaptive multimodal integration.

Yijun Guo, Changsheng You, Changchuan Yin et al.

The performance optimization of UAV communication systems requires the joint design of UAV trajectory and communication efficiently. To tackle the challenge of infinite design variables arising from the continuous-time UAV trajectory optimization, a commonly adopted approach is by approximating the UAV trajectory with piecewise-linear path segments in three-dimensional (3D) space. However, this approach may still incur prohibitive computational complexity in practice when the UAV flight period/distance becomes long, as the distance between consecutive waypoints needs to be kept sufficiently small to retain high approximation accuracy. To resolve this fundamental issue, we propose in this paper a new and general framework for UAV trajectory and communication co-design. First, we propose a flexible path discretization scheme that optimizes only a number of selected waypoints (designable waypoints) along the UAV path for complexity reduction, while all the designable and non-designable waypoints are used in calculating the approximated communication utility along the UAV trajectory for ensuring high trajectory discretization accuracy. Next, given any number of designable waypoints, we propose a novel path compression scheme where the UAV 3D path is first decomposed into three one-dimensional (1D) sub-paths and each sub-path is then approximated by superimposing a number of selected basis paths weighted by their corresponding path coefficients, thus further reducing the path design complexity. Finally, we provide a case study on UAV trajectory design for aerial data harvesting from distributed ground sensors, and numerically show that the proposed schemes can significantly reduce the UAV trajectory design complexity yet achieve favorable rate performance as compared to conventional path/time discretization schemes.

Qingqing Wu, Shuowen Zhang, Beixiong Zheng et al.

Intelligent reflecting surface (IRS) is an enabling technology to engineer the radio signal prorogation in wireless networks. By smartly tuning the signal reflection via a large number of low-cost passive reflecting elements, IRS is capable of dynamically altering wireless channels to enhance the communication performance. It is thus expected that the new IRS-aided hybrid wireless network comprising both active and passive components will be highly promising to achieve a sustainable capacity growth cost-effectively in the future. Despite its great potential, IRS faces new challenges to be efficiently integrated into wireless networks, such as reflection optimization, channel estimation, and deployment from communication design perspectives. In this paper, we provide a tutorial overview of IRS-aided wireless communication to address the above issues, and elaborate its reflection and channel models, hardware architecture and practical constraints, as well as various appealing applications in wireless networks. Moreover, we highlight important directions worthy of further investigation in future work.

Yong Zeng, Xiaoli Xu, Shi Jin et al.

Cellular-connected unmanned aerial vehicle (UAV) is a promising technology to unlock the full potential of UAVs in the future. However, how to achieve ubiquitous three-dimensional (3D) communication coverage for the UAVs in the sky is a new challenge. In this paper, we tackle this challenge by a new coverage-aware navigation approach, which exploits the UAV's controllable mobility to design its navigation/trajectory to avoid the cellular BSs' coverage holes while accomplishing their missions. We formulate an UAV trajectory optimization problem to minimize the weighted sum of its mission completion time and expected communication outage duration, and propose a new solution approach based on the technique of deep reinforcement learning (DRL). To further improve the performance, we propose a new framework called simultaneous navigation and radio mapping (SNARM), where the UAV's signal measurement is used not only for training the deep Q network (DQN) directly, but also to create a radio map that is able to predict the outage probabilities at all locations in the area of interest. This thus enables the generation of simulated UAV trajectories and predicting their expected returns, which are then used to further train the DQN via Dyna technique, thus greatly improving the learning efficiency.

Chongwen Huang, Sha Hu, George C. Alexandropoulos et al.

Future wireless networks are expected to evolve towards an intelligent and software reconfigurable paradigm enabling ubiquitous communications between humans and mobile devices. They will be also capable of sensing, controlling, and optimizing the wireless environment to fulfill the visions of low-power, high-throughput, massively-connected, and low-latency communications. A key conceptual enabler that is recently gaining increasing popularity is the Holographic Multiple Input Multiple Output Surface (HMIMOS) that refers to a low-cost transformative wireless planar structure comprising of sub-wavelength metallic or dielectric scattering particles, which is capable of impacting electromagnetic waves according to desired objectives. In this article, we provide an overview of HMIMOS communications by introducing the available hardware architectures for reconfigurable such metasurfaces and their main characteristics, as well as highlighting the opportunities and key challenges in designing HMIMOS-enabled communications.

Jie Xu, Lingjie Duan, Rui Zhang

Conventional wireless security assumes wireless communications are rightful and aims to protect them against malicious eavesdropping and jamming attacks. However, emerging infrastructure-free mobile communication networks are likely to be illegally used (e.g., by criminals or terrorists) but difficult to be monitored, thus imposing new challenges on the public security. To tackle this issue, this article presents a paradigm shift of wireless security to the surveillance and intervention of infrastructure-free suspicious and malicious wireless communications, by exploiting legitimate eavesdropping and jamming jointly. In particular, {\emph{proactive eavesdropping}} (via jamming) is proposed to intercept and decode information from suspicious communication links for the purpose of inferring their intentions and deciding further measures against them. {\emph{Cognitive jamming}} (via eavesdropping) is also proposed so as to disrupt, disable, and even spoof the targeted malicious wireless communications to achieve various intervention tasks.