arXiv subjects

Zhipeng Lin

Publications and source records attributed to Zhipeng Lin.

At least 19 recordsLinked to original sources

arXiv2026

A Parameter-Specific Retrieval and Knowledge-Guided Reasoning Framework for LLM-Based GPSR Optimization in FANETs

Existing Greedy Perimeter Stateless Routing (GPSR)-based protocols for Flying Ad-Hoc Networks (FANETs) struggle to adapt routing parameters, such as hello interval, multi-path number, and greedy forwarding weights, under highly dynamic environments. As an emerging artificial intelligence technology, large language models (LLMs) show potential for intelligent decision-making, providing new opportunities for adaptive adjustment of GPSR parameters to improve network performance. However, applying LLMs to GPSR remains challenging due to irrelevant experience retrieval and the absence of protocol constraints. To address these issues, we propose a Parameter-Specific Multi-Index Retrieval and Knowledge-Guided Reasoning framework for adaptive GPSR optimization (PMKR-GPSR), an LLM-based framework that enables protocol-consistent routing parameter adaptation. We design a parameter-specific multi-index retrieval mechanism to provide LLMs with parameter-relevant experiences while reducing interference from irrelevant information. We further construct a knowledge-guided constraint graph to enforce that the routing parameters satisfy dependency rules and optimization constraints. Simulation results demonstrate that PMKR-GPSR achieves higher packet delivery ratio and lower end-to-end delay under high-mobility FANETs.

cs.NI↗

arXiv2026

Generating Graph-Like Logical Rules for Knowledge Graph Reasoning via Diffusion Models

Logical rules constitute a cornerstone of knowledge graph (KG) reasoning, valued for their interpretability and ability to model relational patterns. However, existing rule mining methods predominantly focus on simple chain-like rules and therefore neglect the richer relational information encoded in graph-like structures, such as cycles and branches. This limitation is further exacerbated by computational bottlenecks caused by the combinatorial explosion of the search space, which is especially challenging for graph-like rules. Meanwhile, generative approaches such as diffusion models, despite their success in other domains, cannot be directly applied to rule mining because their training objectives are not aligned with the goal of learning high-quality rules, and non-differentiable KG rule quality metrics cannot directly guide model optimization. To address these limitations, we propose GRiD, a framework that reformulates graph-like rule discovery as a discrete generative process conditioned on the target relation. GRiD employs a two-phase training strategy. First, supervised pre-training enables GRiD to capture structural priors from subgraphs sampled from the KG meta-graph. Subsequently, reinforcement learning is applied to fine-tune GRiD through policy gradient optimization guided directly by non-differentiable rule-quality metrics. Experiments on six benchmark datasets show that GRiD achieves competitive performance on KG completion tasks. Ablation studies confirm the efficiency and robustness of GRiD and further show that graph-like rules complement chain-like rules in KG completion. Our code and datasets are available in https://github.com/Haoxiang-Cheng/GRiD.

cs.AI↗

arXiv2025

UAV-Aided Progressive Interference Source Localization Based on Improved Trust Region Optimization

Trust region optimization-based received signal strength indicator (RSSI) interference source localization methods have been widely used in low-altitude research. However, these methods often converge to local optima in complex environments, degrading the positioning performance. This paper presents a novel unmanned aerial vehicle (UAV)-aided progressive interference source localization method based on improved trust region optimization. By combining the Levenberg-Marquardt (LM) algorithm with particle swarm optimization (PSO), our proposed method can effectively enhance the success rate of localization. We also propose a confidence quantification approach based on the UAV-to-ground channel model. This approach considers the surrounding environmental information of the sampling points and dynamically adjusts the weight of the sampling data during the data fusion. As a result, the overall positioning accuracy can be significantly improved. Experimental results demonstrate the proposed method can achieve high-precision interference source localization in noisy and interference-prone environments.

eess.SP↗

arXiv2025

Mitigating Sensitive Information Leakage in LLMs4Code through Machine Unlearning

Large Language Models for Code (LLMs4Code) have achieved strong performance in code generation, but recent studies reveal that they may memorize and leak sensitive information contained in training data, posing serious privacy risks. To address this gap, this work presents the first comprehensive empirical study on applying machine unlearning to mitigate sensitive information leakage in LLMs4Code. We first construct a dedicated benchmark that includes: (i) a synthetic forget set containing diverse forms of personal information, and (ii) a retain set designed to evaluate whether code-generation capability is preserved after unlearning. Using this benchmark, we systematically assess three representative unlearning algorithms (GA, GA+GD, GA+KL) across three widely used open-source LLMs4Code models (AIXCoder-7B, CodeLlama-7B, CodeQwen-7B). Experimental results demonstrate that machine unlearning can substantially reduce direct memorization-based leakage: on average, the direct leak rate drops by more than 50% while retaining about over 91% of the original code-generation performance. Moreover, by analyzing post-unlearning outputs, we uncover a consistent shift from direct to indirect leakage, revealing an underexplored vulnerability that persists even when the target data has been successfully forgotten. Our findings show that machine unlearning is a feasible and effective solution for enhancing privacy protection in LLMs4Code, while also highlighting the need for future techniques capable of mitigating both direct and indirect leakage simultaneously.

cs.CR↗

arXiv2024

Fiber Transmission Model with Parameterized Inputs based on GPT-PINN Neural Network

In this manuscript, a novelty principle driven fiber transmission model for short-distance transmission with parameterized inputs is put forward. By taking into the account of the previously proposed principle driven fiber model, the reduced basis expansion method and transforming the parameterized inputs into parameterized coefficients of the Nonlinear Schrodinger Equations, universal solutions with respect to inputs corresponding to different bit rates can all be obtained without the need of re-training the whole model. This model, once adopted, can have prominent advantages in both computation efficiency and physical background. Besides, this model can still be effectively trained without the needs of transmitted signals collected in advance. Tasks of on-off keying signals with bit rates ranging from 2Gbps to 50Gbps are adopted to demonstrate the fidelity of the model.

cs.AI↗

arXiv2024

Principle Driven Parameterized Fiber Model based on GPT-PINN Neural Network

In cater the need of Beyond 5G communications, large numbers of data driven artificial intelligence based fiber models has been put forward as to utilize artificial intelligence's regression ability to predict pulse evolution in fiber transmission at a much faster speed compared with the traditional split step Fourier method. In order to increase the physical interpretabiliy, principle driven fiber models have been proposed which inserts the Nonlinear Schodinger Equation into their loss functions. However, regardless of either principle driven or data driven models, they need to be re-trained the whole model under different transmission conditions. Unfortunately, this situation can be unavoidable when conducting the fiber communication optimization work. If the scale of different transmission conditions is large, then the whole model needs to be retrained large numbers of time with relatively large scale of parameters which may consume higher time costs. Computing efficiency will be dragged down as well. In order to address this problem, we propose the principle driven parameterized fiber model in this manuscript. This model breaks down the predicted NLSE solution with respect to one set of transmission condition into the linear combination of several eigen solutions which were outputted by each pre-trained principle driven fiber model via the reduced basis method. Therefore, the model can greatly alleviate the heavy burden of re-training since only the linear combination coefficients need to be found when changing the transmission condition. Not only strong physical interpretability can the model posses, but also higher computing efficiency can be obtained. Under the demonstration, the model's computational complexity is 0.0113% of split step Fourier method and 1% of the previously proposed principle driven fiber model.

cs.AI↗

arXiv2024

A Reconfigurable Subarray Architecture and Hybrid Beamforming for Millimeter-Wave Dual-Function-Radar-Communication Systems

Dual-function-radar-communication (DFRC) is a promising candidate technology for next-generation networks. By integrating hybrid analog-digital (HAD) beamforming into a multi-user millimeter-wave (mmWave) DFRC system, we design a new reconfigurable subarray (RS) architecture and jointly optimize the HAD beamforming to maximize the communication sum-rate and ensure a prescribed signal-to-clutter-plus-noise ratio for radar sensing. Considering the non-convexity of this problem arising from multiplicative coupling of the analog and digital beamforming, we convert the sum-rate maximization into an equivalent weighted mean-square error minimization and apply penalty dual decomposition to decouple the analog and digital beamforming. Specifically, a second-order cone program is first constructed to optimize the fully digital counterpart of the HAD beamforming. Then, the sparsity of the RS architecture is exploited to obtain a low-complexity solution for the HAD beamforming. The convergence and complexity analyses of our algorithm are carried out under the RS architecture. Simulations corroborate that, with the RS architecture, DFRC offers effective communication and sensing and improves energy efficiency by 83.4% and 114.2% with a moderate number of radio frequency chains and phase shifters, compared to the persistently- and fullyconnected architectures, respectively.

eess.SP↗

arXiv2023

Sparse Bayesian Learning-Based 3D Spectrum Environment Map Construction-Sampling Optimization, Scenario-Dependent Dictionary Construction and Sparse Recovery

The spectrum environment map (SEM), which can visualize the information of invisible electromagnetic spectrum, is vital for monitoring, management, and security of spectrum resources in cognitive radio (CR) networks. In view of a limited number of spectrum sensors and constrained sampling time, this paper presents a new three-dimensional (3D) SEM construction scheme based on sparse Bayesian learning (SBL). Firstly, we construct a scenario-dependent channel dictionary matrix by considering the propagation characteristic of the interested scenario. To improve sampling efficiency, a maximum mutual information (MMI)-based optimization algorithm is developed for the layout of sampling sensors. Then, a maximum and minimum distance (MMD) clustering-based SBL algorithm is proposed to recover the spectrum data at the unsampled positions and construct the whole 3D SEM. We finally use the simulation data of the campus scenario to construct the 3D SEMs and compare the proposed method with the state-of-the-art. The recovery performance and the impact of different sparsity on the constructed SEMs are also analyzed. Numerical results show that the proposed scheme can reduce the required spectrum sensor number and has higher accuracy under the low sampling rate.

eess.SP↗

arXiv2022

Geometry-Based Stochastic Probability Models for the LoS and NLoS Paths of A2G Channels under Urban Scenario

Path probability prediction is essential to describe the dynamic birth and death of propagation paths, and build the accurate channel model for air-to-ground (A2G) communications. The occurrence probability of each path is complex and time-variant due to fast changeable altitudes of UAVs and scattering environments. Considering the A2G channels under urban scenarios, this paper presents three novel stochastic probability models for the line-of-sight (LoS) path, ground specular (GS) path, and building-scattering (BS) path, respectively. By analyzing the geometric stochastic information of three-dimensional (3D) scattering environments, the proposed models are derived with respect to the width, height, and distribution of buildings. The effect of Fresnel zone and altitudes of transceivers are also taken into account. Simulation results show that the proposed LoS path probability model has good performance at different frequencies and altitudes, and is also consistent with existing models at the low or high altitude. Moreover, the proposed LoS and NLoS path probability models show good agreement with the ray-tracing (RT) simulation method.

eess.SP↗

arXiv2022

Energy-Delay Minimization of Task Migration Based on Game Theory in MEC-assisted Vehicular Networks

Roadside units (RSUs), which have strong computing capability and are close to vehicle nodes, have been widely used to process delay- and computation-intensive tasks of vehicle nodes. However, due to their high mobility, vehicles may drive out of the coverage of RSUs before receiving the task processing results. In this paper, we propose a mobile edge computing-assisted vehicular network, where vehicles can offload their tasks to a nearby vehicle via a vehicle-to-vehicle (V2V) link or a nearby RSU via a vehicle-to-infrastructure link. These tasks are also migrated by a V2V link or an infrastructure-to-infrastructure (I2I) link to avoid the scenario where the vehicles cannot receive the processed task from the RSUs. Considering mutual interference from the same link of offloading tasks and migrating tasks, we construct a vehicle offloading decision-based game to minimize the computation overhead. We prove that the game can always achieve Nash equilibrium and convergence by exploiting the finite improvement property. We then propose a task migration (TM) algorithm that includes three task-processing methods and two task-migration methods. Based on the TM algorithm, computation overhead minimization offloading (COMO) algorithm is presented. Extensive simulation results show that the proposed TM and COMO algorithms reduce the computation overhead and increase the success rate of task processing.

cs.NI↗

arXiv2021

Doppler Shift Estimation in 5G New Radio Non-Terrestrial Networks

Evolving 5G New Radio (NR) to support non-terrestrial networks (NTNs), particularly satellite communication networks, is under exploration in 3GPP. The movement of the spaceborne platforms in NTNs may result in large timing varying Doppler shift that differs for devices in different locations. Using orthogonal frequency-division multiple access (OFDMA) in the uplink, each device will need to apply a different frequency adjustment value to compensate for the Doppler shift. To this end, the 3GPP Release-17 work on NTNs assumes that an NTN device is equipped with a global navigation satellite system (GNSS) chipset and thereby can determine its position and calculate the needed frequency adjustment value using its position information and satellite ephemeris data. This makes GNSS support essential for the NTN operation. However, GNSS signals are weak, not ubiquitous, and susceptible to interference and spoofing. We show that devices without access to GNSS signals can utilize reference signals in more than one frequency position in an OFDM carrier to estimate the Doppler shift and thereby determine the needed frequency adjustment value for pre-compensating the Doppler shift in the uplink. We analyze the performance, elaborate on how to utilize the NR reference signals, and present simulation results. The solution can reduce the dependency of NTN operation on GNSS with reasonable complexity and performance trade-off.

cs.NI↗

arXiv2021

Sum-Rate Maximization for Multi-Reconfigurable Intelligent Surface-Assisted Device-to-Device Communications

This paper proposes to deploy multiple reconfigurable intelligent surfaces (RISs) in device-to-device (D2D)-underlaid cellular systems. The uplink sum-rate of the system is maximized by jointly optimizing the transmit powers of the users, the pairing of the cellular users (CUs) and D2D links, the receive beamforming of the base station (BS), and the configuration of the RISs, subject to the power limits and quality-of-service (QoS) of the users. To address the non-convexity of this problem, we develop a new block coordinate descent (BCD) framework which decouples the D2D-CU pairing, power allocation and receive beamforming, from the configuration of the RISs. Specifically, we derive closed-form expressions for the power allocation and receive beamforming under any D2D-CU pairing, which facilitates interpreting the D2D-CU pairing as a bipartite graph matching solved using the Hungarian algorithm. We transform the configuration of the RISs into a quadratically constrained quadratic program (QCQP) with multiple quadratic constraints. A low-complexity algorithm, named Riemannian manifold-based alternating direction method of multipliers (RM-ADMM), is developed to decompose the QCQP into simpler QCQPs with a single constraint each, and solve them efficiently in a decentralized manner. Simulations show that the proposed algorithm can significantly improve the sum-rate of the D2D-underlaid system with a reduced complexity, as compared to its alternative based on semidefinite relaxation (SDR).

cs.IT↗

arXiv2021

Few-Sample Named Entity Recognition for Security Vulnerability Reports by Fine-Tuning Pre-Trained Language Models

Public security vulnerability reports (e.g., CVE reports) play an important role in the maintenance of computer and network systems. Security companies and administrators rely on information from these reports to prioritize tasks on developing and deploying patches to their customers. Since these reports are unstructured texts, automatic information extraction (IE) can help scale up the processing by converting the unstructured reports to structured forms, e.g., software names and versions and vulnerability types. Existing works on automated IE for security vulnerability reports often rely on a large number of labeled training samples. However, creating massive labeled training set is both expensive and time consuming. In this work, for the first time, we propose to investigate this problem where only a small number of labeled training samples are available. In particular, we investigate the performance of fine-tuning several state-of-the-art pre-trained language models on our small training dataset. The results show that with pre-trained language models and carefully tuned hyperparameters, we have reached or slightly outperformed the state-of-the-art system on this task. Consistent with previous two-step process of first fine-tuning on main category and then transfer learning to others as in [7], if otherwise following our proposed approach, the number of required labeled samples substantially decrease in both stages: 90% reduction in fine-tuning from 5758 to 576,and 88.8% reduction in transfer learning with 64 labeled samples per category. Our experiments thus demonstrate the effectiveness of few-sample learning on NER for security vulnerability report. This result opens up multiple research opportunities for few-sample learning for security vulnerability reports, which is discussed in the paper. Code: https://github.com/guanqun-yang/FewVulnerability.

cs.CL↗

arXiv2021

Wide-Beam Array Antenna Power Gain Maximization via ADMM Framework

This paper proposes two algorithms to maximize the minimum array power gain in a wide-beam mainlobe by solving the power gain pattern synthesis (PGPS) problem with and without sidelobe constraints. Firstly, the nonconvex PGPS problem is transformed into a nonconvex linear inequality optimization problem and then converted to an augmented Lagrangian problem by introducing auxiliary variables via the Alternating Direction Method of Multipliers (ADMM) framework. Next,the original intractable problem is converted into a series of nonconvex and convex subproblems. The nonconvex subproblems are solved by dividing their solution space into a finite set of smaller ones, in which the solution would be obtained pseudoanalytically. In such a way, the proposed algorithms are superior to the existing PGPS-based ones as their convergence can be theoretically guaranteed with a lower computational burden. Numerical examples with both isotropic element pattern (IEP) and active element pattern (AEP) arrays are simulated to show the effectiveness and superiority of the proposed algorithms by comparing with the related existing algorithms.

eess.SP↗

arXiv2021

Energy-Efficient Resource Allocation in Massive MIMO-NOMA Networks with Wireless Power Transfer: A Distributed ADMM Approach

In multicell massive multiple-input multiple-output (MIMO) non-orthogonal multiple access (NOMA) networks, base stations (BSs) with multiple antennas deliver their radio frequency energy in the downlink, and Internet-of-Things (IoT) devices use their harvested energy to support uplink data transmission. This paper investigates the energy efficiency (EE) problem for multicell massive MIMO NOMA networks with wireless power transfer (WPT). To maximize the EE of the network, we propose a novel joint power, time, antenna selection, and subcarrier resource allocation scheme, which can properly allocate the time for energy harvesting and data transmission. Both perfect and imperfect channel state information (CSI) are considered, and their corresponding EE performance is analyzed. Under quality-of-service (QoS) requirements, an EE maximization problem is formulated, which is non-trivial due to non-convexity. We first adopt nonlinear fraction programming methods to convert the problem to be convex, and then, develop a distributed alternating direction method of multipliers (ADMM)- based approach to solve the problem. Simulation results demonstrate that compared to alternative methods, the proposed algorithm can converge quickly within fewer iterations, and can achieve better EE performance.

cs.IT↗

arXiv2021

Joint Estimation of Multipath Angles and Delays for Millimeter-Wave Cylindrical Arrays with Hybrid Front-ends

Accurate channel parameter estimation is challenging for wideband millimeter-wave (mmWave) large-scale hybrid arrays, due to beam squint and much fewer radio frequency (RF) chains than antennas. This paper presents a novel joint delay and angle estimation approach for wideband mmWave fully-connected hybrid uniform cylindrical arrays. We first design a new hybrid beamformer to reduce the dimension of received signals on the horizontal plane by exploiting the convergence of the Bessel function, and to reduce the active beams in the vertical direction through preselection. The important recurrence relationship of the received signals needed for subspace-based angle and delay estimation is preserved, even with substantially fewer RF chains than antennas. Then, linear interpolation is generalized to reconstruct the received signals of the hybrid beamformer, so that the signals can be coherently combined across the whole band to suppress the beam squint. As a result, efficient subspace-based algorithm algorithms can be developed to estimate the angles and delays of multipath components. The estimated delays and angles are further matched and correctly associated with different paths in the presence of non-negligible noises, by putting forth perturbation operations. Simulations show that the proposed approach can approach the Cram\'{e}r-Rao lower bound (CRLB) of the estimation with a significantly lower computational complexity than existing techniques.

cs.IT↗

arXiv2020

Tensor-based Multi-dimensional Wideband Channel Estimation for mmWave Hybrid Cylindrical Arrays

Channel estimation is challenging for hybrid millimeter wave (mmWave) large-scale antenna arrays which are promising in 5G/B5G applications. The challenges are associated with angular resolution losses resulting from hybrid front-ends, beam squinting, and susceptibility to the receiver noises. Based on tensor signal processing, this paper presents a novel multi-dimensional approach to channel parameter estimation with large-scale mmWave hybrid uniform circular cylindrical arrays (UCyAs) which are compact in size and immune to mutual coupling but known to suffer from infinite-dimensional array responses and intractability. We design a new resolution-preserving hybrid beamformer and a low-complexity beam squinting suppression method, and reveal the existence of shift-invariance relations in the tensor models of received array signals at the UCyA. Exploiting these relations, we propose a new tensor-based subspace estimation algorithm to suppress the receiver noises in all dimensions (time, frequency, and space). The algorithm can accurately estimate the channel parameters from both coherent and incoherent signals. Corroborated by the Cram\'{e}r-Rao lower bound (CRLB), simulation results show that the proposed algorithm is able to achieve substantially higher estimation accuracy than existing matrix-based techniques, with a comparable computational complexity.

eess.SP↗

arXiv2020

Complex ResNet Aided DoA Estimation for Near-Field MIMO Systems

The near-field effect of short-range multiple-input multiple-output (MIMO) systems imposes many challenges on direction-of-arrival (DoA) estimation. Most conventional scenarios assume that the far-field planar wavefronts hold. In this paper, we investigate the DoA estimation problem in short-range MIMO communications, where the effect of near-field spherical wave is non-negligible. By converting it into a regression task, a novel DoA estimation framework based on complex-valued deep learning (CVDL) is proposed for the near-field region in short-range MIMO communication systems. Under the assumption of a spherical wave model, the array steering vector is determined by both the distance and the direction. However, solving this regression task containing a massive number of variables is challenging, since datasets need to capture numerous complicated feature representations. To overcome this, a virtual covariance matrix (VCM) based on received signals is constructed, and thus such features extracted from the VCM can deal with the complicated coupling relationship between the direction and the distance. Although the emergence of wireless big data driven by future communication networks promotes deep learning-based wireless signal processing, the learning algorithms of complex-valued signals are still ongoing. This paper proposes a one-dimensional (1-D) residual network that can directly tackle complex-valued features due to the inherent 1-D structure of signal subspace vectors. In addition, we put forth a cropped VCM based policy which can be applied to different antenna sizes. The proposed method is able to fully exploit the complex-valued information. Our simulation results demonstrate the superiority of the proposed CVDL approach over the baseline schemes in terms of the accuracy of DoA estimation.

eess.SP↗