arXiv subjects

Xiaoheng Deng

Publications and source records attributed to Xiaoheng Deng.

14 recordsLinked to original sources

arXiv2026

Domain-Adaptive Deep Joint Source-Channel Coding for Image Classification

Deep joint source--channel coding (Deep JSCC) enables visual semantic transmission by mapping inputs directly to channel symbols and task outputs, but its performance can deteriorate under distribution shifts between training and deployment domains. We study single-source domain adaptation for task-oriented Deep JSCC and formulate a classification-capacity-invariance (CCI) function to characterize how the available channel capacity and class-conditional cross-domain invariance affect target domain classification accuracy. A scalar linear analysis of source-domain-optimal solutions and a controlled shallow nonlinear validation show that target domain classification accuracy can vary non-monotonically with the invariance constraint and with available capacity along separate control paths obtained by varying the transmitted dimension or CSNR. We then propose a domain-adaptive Deep JSCC framework that combines pseudo-label-based class-level adversarial alignment with supervised contrastive learning on confidence-filtered target samples. Experiments on digit and PACS datasets over AWGN and Rayleigh fading channels demonstrate improved target domain generalization without introducing additional inference-time networks. On SVHN $\rightarrow$ MNIST, the proposed method achieves 98.15\% target-domain accuracy at a CSNR of 10 dB.

cs.IT↗

arXiv2026

SpaceVLN: A Zero-Shot Vision-and-Language Navigation Agent with Online Spatial Cognitive Memory and Reasoning

Vision-and-Language Navigation in continuous environments requires agents to understand the spatial structure of previously unseen environments in order to follow language instructions. Although foundation models have opened a promising path toward zero-shot navigation without task-specific policy training, many navigators still rely on local visual cues and linear history-based reasoning, overlooking the spatial nature of navigation across explored regions, traversed paths, landmarks, and their spatial relations. In this paper, we propose SpaceVLN, a navigation agent built around Spatial Cognitive Memory and Task-Guided Spatial Reasoning. Specifically, SpaceVLN introduces an efficient stagewise closed-loop framework where planning and execution are organized around verifiable space--landmark stages. During navigation, the agent progressively abstracts explored regions into Spatial Waypoints and dynamically maintains subtask-grounded landmark evidence, forming a hierarchical Spatial Cognitive Memory for progress localization and spatial-relation understanding. Built on this memory, Spatial-CoT integrates task-progress reasoning with spatial perception, analysis, and prediction, enabling Task-Guided Spatial Reasoning for embodied navigation. The unified stage interface enables SpaceVLN to address both Vision-and-Language Navigation and Object-Goal Navigation under a unified zero-shot setting, without task-specific policy training. Across R2R-CE, RxR-CE, GN-Bench, and HM3D-OVON, SpaceVLN achieves state-of-the-art zero-shot performance, and real-robot deployment further validates its applicability. These results highlight Spatial Cognitive Memory and Task-Guided Spatial Reasoning as a practical foundation for stronger embodied navigation agents.

cs.RO↗

arXiv2026

Is VLA Reasoning Faithful? Probing Safety of Chain-of-Causation in Autonomous Driving Models

We present the first systematic study of faithfulness in Vision-Language-Action (VLA) driving models, analyzing 300 Alpamayo-R1-10B inferences across 100 diverse PhysicalAI-AV scenarios. Our main finding is that output natural-language rationales with trajectories may be significantly unfaithful: (i) overall reasoning fidelity is only 42.5%, with Chain-of-Causation matching scene reality less than half the time; (ii) 94 missed pedestrians in one-third of pedestrian-relevant scenes; (iii) 97.7% trajectory fragility under mild visual perturbations; and (iv) only 48.3% mean reasoning-action consistency, with 53.3% of inferences exhibiting low consistency, including 37.9% of stop-claimed cases where the model continues instead. We formalize faithfulness information-theoretically, define entity and action fidelity with verification criteria, and outline a four-component safety architecture aligned with these results.

cs.AI↗

arXiv2026

Resilient and Freshness-Aware Scheduling for Industrial Multi-Hop IAB Networks: A Packet Duplication Approach

In industrial millimeter-wave (mmWave) multi-hop Integrated Access and Backhaul (IAB) networks, dynamic blockages caused by moving obstacles pose a severe threat to robust and continuous networks. While Packet Duplication (PD) enhances reliability by path diversity, it inevitably doubles the traffic load, leading to severe congestion and degraded Age of Information (AoI). To navigate this reliability-congestion trade-off, we formulated an optimization problem in a multi-hop IAB scenario that minimizes the average AOI while satisfying strict queue stability constraints. We utilize Lyapunov optimization to transform the long-term stochastic optimization problem into tractable deterministic sub-problems. To solve these sub-problems efficiently, we propose a Resilient and Freshness-Aware Scheduling (RFAS) algorithm. Simulation results show that in blockage-prone environments, RFAS significantly outperforms baselines by maintaining a Packet Delivery Ratio (PDR) above 95\%. Crucially, it strictly guarantees queue stability under hard buffer constraints, whereas baselines suffer from buffer overflows. Furthermore, RFAS reduces the network load imbalance by 19\% compared to the baseline in high-frequency traffic scenarios. This confirms RFAS as a robust and sustainable solution for real-time industrial control loops.

cs.NI↗

arXiv2025

AoI-Energy-Spectrum Optimization in Post-Disaster Powered Communication Intelligent Network via Hierarchical Heterogeneous Graph Neural Network

This paper designs a post-disaster powered communication intelligent network (PDPCIN) to address communication disruptions caused by ground base station (GBS) failures within the post-disaster area. PDPCIN employs unmanned aerial vehicles (UAVs) to provide wireless data collection (WDC) and wireless energy transmission (WET) for affected areas and leverages low earth orbit satellites (LEO SATs) to relay UAV data to the nearest survival GBS. To ensure basic post-disaster communication while co-optimizing age of information (AoI), energy efficiency, and spectrum efficiency, intelligent synchronization-UAV (IS-UAV) architecture, AoI-based four thresholds updating (AFTU) mechanism, and Dynamic multi-LEO access (DMLA) strategy are proposed. However, three key challenges remain: time-varying task-resource imbalances, complex topology caused by multi-device scheduling, and nonlinear coupling in multidimensional metric optimization, making system optimization NP-hard. Therefore, this paper proposes a hierarchical heterogeneous graph neural networks (HHGNN) framework. It models heterogeneous device nodes and their communication relations as a hierarchical heterogeneous graph structure, integrating our defined graph sensing, exchange, and mask layer to handle the network's input, feature propagation, and output. To search appropriate number of single-LEO SATs, we propose single-LEO SAT density optimization (S-LSDO) algorithm. Finally, we compare the proposed scheme with state-of-the-art benchmarks to validate its superior collaborative optimization of AoI, energy efficiency, and spectrum efficiency. Based on this, we derive the expressions for the expected values of AoI and stagnant AoI proportion.

cs.NI↗

arXiv2024

MetaDD: Boosting Dataset Distillation with Neural Network Architecture-Invariant Generalization

Dataset distillation (DD) entails creating a refined, compact distilled dataset from a large-scale dataset to facilitate efficient training. A significant challenge in DD is the dependency between the distilled dataset and the neural network (NN) architecture used. Training a different NN architecture with a distilled dataset distilled using a specific architecture often results in diminished trainning performance for other architectures. This paper introduces MetaDD, designed to enhance the generalizability of DD across various NN architectures. Specifically, MetaDD partitions distilled data into meta features (i.e., the data's common characteristics that remain consistent across different NN architectures) and heterogeneous features (i.e., the data's unique feature to each NN architecture). Then, MetaDD employs an architecture-invariant loss function for multi-architecture feature alignment, which increases meta features and reduces heterogeneous features in distilled data. As a low-memory consumption component, MetaDD can be seamlessly integrated into any DD methodology. Experimental results demonstrate that MetaDD significantly improves performance across various DD methods. On the Distilled Tiny-Imagenet with Sre2L (50 IPC), MetaDD achieves cross-architecture NN accuracy of up to 30.1\%, surpassing the second-best method (GLaD) by 1.7\%.

cs.CV↗

arXiv2024

Fully Exploiting Every Real Sample: SuperPixel Sample Gradient Model Stealing

Model stealing (MS) involves querying and observing the output of a machine learning model to steal its capabilities. The quality of queried data is crucial, yet obtaining a large amount of real data for MS is often challenging. Recent works have reduced reliance on real data by using generative models. However, when high-dimensional query data is required, these methods are impractical due to the high costs of querying and the risk of model collapse. In this work, we propose using sample gradients (SG) to enhance the utility of each real sample, as SG provides crucial guidance on the decision boundaries of the victim model. However, utilizing SG in the model stealing scenario faces two challenges: 1. Pixel-level gradient estimation requires extensive query volume and is susceptible to defenses. 2. The estimation of sample gradients has a significant variance. This paper proposes Superpixel Sample Gradient stealing (SPSG) for model stealing under the constraint of limited real samples. With the basic idea of imitating the victim model's low-variance patch-level gradients instead of pixel-level gradients, SPSG achieves efficient sample gradient estimation through two steps. First, we perform patch-wise perturbations on query images to estimate the average gradient in different regions of the image. Then, we filter the gradients through a threshold strategy to reduce variance. Exhaustive experiments demonstrate that, with the same number of real samples, SPSG achieves accuracy, agreements, and adversarial success rate significantly surpassing the current state-of-the-art MS methods. Codes are available at https://github.com/zyl123456aB/SPSG_attack.

cs.CV↗

arXiv2024

Haina Storage: A Decentralized Secure Storage Framework Based on Improved Blockchain Structure

Although the decentralized storage technology based on the blockchain can effectively realize secure data storage on cloud services. However, there are still some problems in the existing schemes, such as low storage capacity and low efficiency. To address related issues, we propose a novel decentralized storage framework, which mainly includes four aspects: (1) we proposed a Bi-direction Circular Linked Chain Structure (BCLCS), which improves data's storage capacity and applicability in decentralized storage. (2) A Proof of Resources (PoR) decision model is proposed. By introducing the network environment as an essential evaluation parameter of storage right decision, the energy and time consumption of decision-making are reduced, and the fairness of decision-making is improved. (3) A chain structure dynamic locking mechanism (CSDLM) is designed to realize anti-traverse and access control. (4) A Bi-directional data Access Mechanism (BDAM) is proposed, which improves the efficiency of data access and acquisition in decentralized storage mode. The experimental results show that the framework has significantly improved the shortcomings of the current decentralized storage.

cs.CR↗

arXiv2023

Dependent Task Offloading in Edge Computing Using GNN and Deep Reinforcement Learning

Task offloading is a widely used technology in Mobile Edge Computing (MEC), which declines the completion time of user task with the help of resourceful edge servers. Existing works mainly focus on the case that the computation density of a user task is homogenous so that it can be offloaded in full or by percentage. However, various user tasks in real life consist of several inner dependent subtasks, each of which is a minimum execution unit logically. Motivated by this gap, we aim to solve the Dependent Task Offloading (DTO) problem under multi-user multi-edge scenario in this paper. We firstly use Directed Acyclic Graph (DAG) to represent dependent task where nodes indicate subtasks and directed edges indicate dependencies among subtasks. Then we propose a scheme based on Graph Attention Network (GAT) and Deep Reinforcement Learning (DRL) to minimize the makespan of user tasks. To utilize GAT efficiently, we put the training of it on resourceful cloud in unsupervised style due to the numerous data and computation resource requirements. In addition, we design a multi-discrete Action space for DRL algorithm to enhance the applicability of our proposed scheme. Experiments are conducted on broadly distributed synthetic data. The results demonstrate that our proposed approach can be adapted to both simple and complex MEC environments and outperforms other methods.

cs.DC↗

arXiv2022

Actions at the Edge: Jointly Optimizing the Resources in Multi-access Edge Computing

Multi-access edge computing (MEC) is an emerging paradigm that pushes resources for sensing, communications, computing, storage and intelligence (SCCSI) to the premises closer to the end users, i.e., the edge, so that they could leverage the nearby rich resources to improve their quality of experience (QoE). Due to the growing emerging applications targeting at intelligentizing life-sustaining cyber-physical systems, this paradigm has become a hot research topic, particularly when MEC is utilized to provide edge intelligence and real-time processing and control. This article is to elaborate the research issues along this line, including basic concepts and performance metrics, killer applications, architectural design, modeling approaches and solutions, and future research directions. It is hoped that this article provides a quick introduction to this fruitful research area particularly for beginning researchers.

cs.NI↗

arXiv2019

UAV-aided urban target tracking system based on edge computing

Target tracking is an important issue of social security. In order to track a target, traditionally a large amount of surveillance video data need to be uploaded into the cloud for processing and analysis, which put stremendous bandwidth pressure on communication links in access networks and core networks. At the same time, the long delay in wide area network is very likely to cause a tracking system to lose its target. Often, unmanned aerial vehicle (UAV) has been adopted for target tracking due to its flexibility, but its limited flight time due to battery constraint and the blocking by various obstacles in the field pose two major challenges to its target tracking task, which also very likely results in the loss of target. A novel target tracking model that coordinates the tracking by UAV and ground nodes in an edge computing environment is proposed in this study. The model can effectively reduce the communication cost and the long delay of the traditional surveillance camera system that relies on cloud computing, and it can improve the probability of finding a target again after an UAV loses the tracing of that target. It has been demonstrated that the proposed system achieved a significantly better performance in terms of low latency, high reliability, and optimal quality of experience (QoE).

eess.SP↗

arXiv2011

Degree and connectivity of the Internet's scale-free topology

In this paper we theoretically and empirically study the degree and connectivity of the Internet's scale-free topology at the autonomous system (AS) level. The basic features of the scale-free network have influence on the normalization constant of the degree distribution p(k). We develop a mathematics model of the Internet's scale-free topology. On this model we theoretically get the formulas of the average degree, the ratios of the kmin-degree (minimum degree) nodes and the kmax-degree (maximum degree) nodes, the fraction of the degrees (or links) in the hands of the richer (top best-connected) nodes. We find the average degree is larger for smaller power-law exponent λ and larger minimum or maximum degree. The ratio of the kmin-degree nodes is larger for larger λ and smaller kmin or kmax. The ratio of the kmax-degree ones is larger for smaller λ and kmax or larger kmin. The richer nodes hold most of the total degrees of the AS-level Internet topology. In addition, we reveal the ratio of the kmin-degree nodes or the rate of the increase of the average degree has power-law decay with the increase of the kmin. The ratio of the kmax-degree nodes has power-law decay with the increase of the kmax, and the fraction of the degrees in the hands of the richer 27% nodes is about 73% (the '73/27 rule'). At last, we empirically calculate, based on empirical data extracted from BGP, the average degree and the ratio and fraction using our method and other methods, and find that our method is rigorous and effective for the AS-level Internet topology.

cs.NI↗

arXiv2011

Modelling on the Guaranteed QoS for Wireless Sensor Networks: A Network Calculus Approach

Wireless sensor networks (WSNs) became one of the high technology domains during the last ten years. Real-time applications for them make it necessary to provide the guaranteed Quality of Service (QoS). The main contributions of this paper are a system skeleton and a guaranteed QoS model that are suitable for the WSNs. To do it, we develop a sensor node model based on virtual buffer sharing and present a two-layer scheduling model using the network calculus. With the system skeleton, we develop a guaranteed QoS model, such as the upper bounds on buffer queue length/delay/effective bandwidth, and single-hop/ multi-hops delay/jitter/effective bandwidth. Numerical results show the system skeleton and the guaranteed QoS model are scalable for different types of flows, including the self-similar traffic flows, and the parameters of flow regulators and service curves of sensor nodes affect them. Our proposal leads to buffer dimensioning, guaranteed QoS support and control in the WSNs.

cs.NI↗

arXiv2005

MAEC : A Movement-Assisted Energy Conserving Method in Event Driven Wireless Sensor Networks

Energy is one of the most important resources in wireless sensor networks. Recently, the mobility of base station has been exploited to preserve the energy. But in event driven networks, the mobility issue is quite different from the continuous monitoring one because only a small portion of sensor node has data to send at one time. The number of sensor node that forward traffic should be minimized to prolong the network lifetime. In this paper, we propose a movement-assisted energy conserving method which tries to reduce the amount of forwarding sensor node by directing the base station to move close to the hotspots. This method achieves good performance especially when applied to a network with a set of cooperative mobile base station. Extensive simulation has been done to verify the effectiveness of the propose schema.

cs.NI↗