arXiv subjects

Zhaoxin Zhang

Publications and source records attributed to Zhaoxin Zhang.

12 recordsLinked to original sources

arXiv2026

DoHFuse: A Dual-Branch Architecture with DMAGLSTM for Website Fingerprinting over DNS over HTTPS/3

As personal data privacy becomes increasingly critical in Internet of Things (IoT) environments, secure DNS protocols such as DNS over HTTPS (DoH) and DNS over TLS (DoT) have been widely adopted to protect device communications. However, without effective obfuscation, these protocols remain vulnerable to Website Fingerprinting (WF) attacks that can reveal user activity. With the ongoing deployment of DNS over HTTP/3 (DoH/3) in modern networked systems, padding strategies have been increasingly applied in practice. It is therefore essential to investigate whether DoH/3 can effectively mitigate WF attacks in realistic IoT and edge-network scenarios. To address this, we first collect and publicly release the first real-world benchmark dataset of DoH/3 traffic, generated from domain resolution processes in practical network environments. We further propose DoHFuse, a dual-branch learning framework that integrates inter-arrival time sequences and refined statistical features through an improved DMAG-LSTM, specifically designed to capture burst-aligned temporal patterns. Experimental results show that DoHFuse achieves an accuracy of 88.05% (precision 88.56, recall 87.96, F1 87.83) in a closed-world setting of 449 classes, and an AUPRC of 0.975 with an F1 score of 0.951 (precision 0.906, recall 1.0) in open-world detection. These findings demonstrate that DoH/3 traffic remains susceptible to WF attacks, suggesting that commonly deployed padding mechanisms alone are insufficient to ensure privacy protection in IoT-scale encrypted DNS communications.

cs.CR↗

arXiv2026

Beyond Viewpoint Generalization: What Multi-View Demonstrations Offer and How to Synthesize Them for Robot Manipulation?

Does multi-view demonstration truly improve robot manipulation, or merely enhance cross-view robustness? We present a systematic study quantifying the performance gains, scaling behavior, and underlying mechanisms of multi-view data for robot manipulation. Controlled experiments show that, under both fixed and randomized backgrounds, multi-view demonstrations consistently improve single-view policy success and generalization. Performance varies non-monotonically with view coverage, revealing effective regimes rather than a simple "more is better" trend. Notably, multi-view data breaks the scaling limitation of single-view datasets and continues to raise performance ceilings after saturation. Mechanistic analysis shows that multi-view learning promotes manipulation-relevant visual representations, better aligns the action head with the learned feature distribution, and reduces overfitting. Motivated by the importance of multi-view data and its scarcity in large-scale robotic datasets, as well as the difficulty of collecting additional viewpoints in real world settings, we propose RoboNVS, a geometry-aware self-supervised framework that synthesizes novel-view videos from monocular inputs. The generated data consistently improves downstream policies in both simulation and real-world environments.

cs.RO↗

arXiv2025

When Harmless Words Harm: A New Threat to LLM Safety via Conceptual Triggers

Recent research on large language model (LLM) jailbreaks has primarily focused on techniques that bypass safety mechanisms to elicit overtly harmful outputs. However, such efforts often overlook attacks that exploit the model's capacity for abstract generalization, creating a critical blind spot in current alignment strategies. This gap enables adversaries to induce objectionable content by subtly manipulating the implicit social values embedded in model outputs. In this paper, we introduce MICM, a novel, model-agnostic jailbreak method that targets the aggregate value structure reflected in LLM responses. Drawing on conceptual morphology theory, MICM encodes specific configurations of nuanced concepts into a fixed prompt template through a predefined set of phrases. These phrases act as conceptual triggers, steering model outputs toward a specific value stance without triggering conventional safety filters. We evaluate MICM across five advanced LLMs, including GPT-4o, Deepseek-R1, and Qwen3-8B. Experimental results show that MICM consistently outperforms state-of-the-art jailbreak techniques, achieving high success rates with minimal rejection. Our findings reveal a critical vulnerability in commercial LLMs: their safety mechanisms remain susceptible to covert manipulation of underlying value alignment.

cs.CL↗

arXiv2025

GPR: Towards a Generative Pre-trained One-Model Paradigm for Large-Scale Advertising Recommendation

As an intelligent infrastructure connecting users with commercial content, advertising recommendation systems play a central role in information flow and value creation within the digital economy. However, existing multi-stage advertising recommendation systems suffer from objective misalignment and error propagation, making it difficult to achieve global optimality, while unified generative recommendation models still struggle to meet the demands of practical industrial applications. To address these issues, we propose GPR (Generative Pre-trained Recommender), the first one-model framework that redefines advertising recommendation as an end-to-end generative task, replacing the traditional cascading paradigm with a unified generative approach. To realize GPR, we introduce three key innovations spanning unified representation, network architecture, and training strategy. First, we design a unified input schema and tokenization method tailored to advertising scenarios, mapping both ads and organic content into a shared multi-level semantic ID space, thereby enhancing semantic alignment and modeling consistency across heterogeneous data. Second, we develop the Heterogeneous Hierarchical Decoder (HHD), a dual-decoder architecture that decouples user intent modeling from ad generation, achieving a balance between training efficiency and inference flexibility while maintaining strong modeling capacity. Finally, we propose a multi-stage joint training strategy that integrates Multi-Token Prediction (MTP), Value-Aware Fine-Tuning and the Hierarchy Enhanced Policy Optimization (HEPO) algorithm, forming a complete generative recommendation pipeline that unifies interest modeling, value alignment, and policy optimization. GPR has been fully deployed in the Tencent Weixin Channels advertising system, delivering significant improvements in key business metrics including GMV and CTCVR.

cs.IR↗

arXiv2025

MG-HGNN: A Heterogeneous GNN Framework for Indoor Wi-Fi Fingerprint-Based Localization

Received signal strength indicator (RSSI) is the primary representation of Wi-Fi fingerprints and serves as a crucial tool for indoor localization. However, existing RSSI-based positioning methods often suffer from reduced accuracy due to environmental complexity and challenges in processing multi-source information. To address these issues, we propose a novel multi-graph heterogeneous GNN framework (MG-HGNN) to enhance spatial awareness and improve positioning performance. In this framework, two graph construction branches perform node and edge embedding, respectively, to generate informative graphs. Subsequently, a heterogeneous graph neural network is employed for graph representation learning, enabling accurate positioning. The MG-HGNN framework introduces the following key innovations: 1) multi-type task-directed graph construction that combines label estimation and feature encoding for richer graph information; 2) a heterogeneous GNN structure that enhances the performance of conventional GNN models. Evaluations on the UJIIndoorLoc and UTSIndoorLoc public datasets demonstrate that MG-HGNN not only achieves superior performance compared to several state-of-the-art methods, but also provides a novel perspective for enhancing GNN-based localization methods. Ablation studies further confirm the rationality and effectiveness of the proposed framework.

cs.LG↗

arXiv2025

HMCGeo: IP Region Prediction Based on Hierarchical Multi-label Classification

Fine-grained IP geolocation plays a critical role in applications such as location-based services and cybersecurity. Most existing fine-grained IP geolocation methods are regression-based; however, due to noise in the input data, these methods typically encounter kilometer-level prediction errors and provide incorrect region information for users. To address this issue, this paper proposes a novel hierarchical multi-label classification framework for IP region prediction, named HMCGeo. This framework treats IP geolocation as a hierarchical multi-label classification problem and employs residual connection-based feature extraction and attention prediction units to predict the target host region across multiple geographical granularities. Furthermore, we introduce probabilistic classification loss during training, combining it with hierarchical cross-entropy loss to form a composite loss function. This approach optimizes predictions by utilizing hierarchical constraints between regions at different granularities. IP region prediction experiments on the New York, Los Angeles, and Shanghai datasets demonstrate that HMCGeo achieves superior performance across all geographical granularities, significantly outperforming existing IP geolocation methods.

cs.LG↗

arXiv2024

A QoE-Aware Split Inference Accelerating Algorithm for NOMA-based Edge Intelligence

Even the AI has been widely used and significantly changed our life, deploying the large AI models on resource limited edge devices directly is not appropriate. Thus, the model split inference is proposed to improve the performance of edge intelligence, in which the AI model is divided into different sub models and the resource-intensive sub model is offloaded to edge server wirelessly for reducing resource requirements and inference latency. However, the previous works mainly concentrate on improving and optimizing the system QoS, ignore the effect of QoE which is another critical item for the users except for QoS. Even the QoE has been widely learned in EC, considering the differences between task offloading in EC and split inference in EI, and the specific issues in QoE which are still not addressed in EC and EI, these algorithms cannot work effectively in edge split inference scenarios. Thus, an effective resource allocation algorithm is proposed in this paper, for accelerating split inference in EI and achieving the tradeoff between inference delay, QoE, and resource consumption, abbreviated as ERA. Specifically, the ERA takes the resource consumption, QoE, and inference latency into account to find the optimal model split strategy and resource allocation strategy. Since the minimum inference delay and resource consumption, and maximum QoE cannot be satisfied simultaneously, the gradient descent based algorithm is adopted to find the optimal tradeoff between them. Moreover, the loop iteration GD approach is developed to reduce the complexity of the GD algorithm caused by parameter discretization. Additionally, the properties of the proposed algorithms are investigated, including convergence, complexity, and approximation error. The experimental results demonstrate that the performance of ERA is much better than that of the previous studies.

cs.LG↗

arXiv2024

6Rover: Leveraging Reinforcement Learning-based Address Pattern Mining Approach for Discovering Active Targets in IPv6 Unseeded Space

The discovery of active IPv6 addresses represents a pivotal challenge in IPv6 network survey, as it is a prerequisite for downstream tasks such as network topology measurements and security analysis. With the rapid spread of IPv6 networks in recent years, many researchers have focused on improving the hit rate, efficiency, and coverage of IPv6 scanning methods, resulting in considerable advancements. However, existing approaches remain heavily dependent on seed addresses, thereby limiting their effectiveness in unseeded prefixes. Consequently, this paper proposes 6Rover, a reinforcement learning-based model for active address discovery in unseeded environments. To overcome the reliance on seeded addresses, 6Rover constructs patterns with higher generality that reflects the actual address allocation strategies of network administrators, thereby avoiding biased transfers of patterns from seeded to unseeded prefixes. After that, 6Rover employs a multi-armed bandit model to optimize the probing resource allocation when applying patterns to unseeded spaces. It models the challenge of discovering optimal patterns in unseeded spaces as an exploration-exploitation dilemma, and progressively uncover the potential patterns applied in unseeded spaces, leading to the efficient discovery of active addresses without seed address as the prior knowledge. Experiments on large-scale unseeded datasets show that 6Rover has a higher hit rate than existing methods in the absence of any seed addresses as prior knowledge. In real network environments, 6Rover achieved a 5% - 8% hit rate in seedless spaces with 100 million budget scale, representing an approximate 200\% improvement over the existing state-of-the-art methods.

cs.NI↗

arXiv2022

CoNet: Borderless and decentralized server cooperation in edge computing

In edge computing (EC), by offloading tasks to edge server or remote cloud, the system performance can be improved greatly. However, since the traffic distribution in EC is heterogeneous and dynamic, it is difficult for an individual edge server to provide satisfactory computation service anytime and anywhere. This issue motivated the researchers to study the cooperation between edge servers. The previous server cooperation algorithms have disadvantages since the cooperated region is limited within one-hop. However, the performance of EC can be improved further by releasing the restriction of cooperation region. Even some works have extended the cooperated region to multi-hops, they fail to support the task offloading which is one of the core issues of edge computing. Therefore, we propose a new decentralized and borderless server cooperation algorithm for edge computing which takes task offloading strategy into account, named CoNet. In CoNet, the cooperation region is not limited. Each server forms its own basic cooperation unit (BCU) and calculates its announced capability based on BCU. The server's capability, the processing delay, the task and calculation result forwarding delay are considered during the calculation. The task division strategy bases on the real capability of host-server and the announced capability of cooperation-servers. This cooperation process is recursive and will be terminated once the terminal condition is satisfied. The simulation results demonstrate the advantages of CoNet over previous works.

cs.NI↗

arXiv2019

Pairwise-based Multi-Attribute Decision Making Approach for Wireless Network

In the wireless network applications, such as routing decision, network selection, etc., the Multi-Attribute Decision Making (MADM) has been widely used. The MADM approach can address the multi-objective decision making issues effectively.However, when the parameters vary greatly, the traditionalMADM algorithm is not effective anymore. To solve this problem,in this paper, we propose the pairwise-based MADM algorithm.In the PMADM, only two nodes utilities are calculated and compared each time. The PMADM algorithm is much more accurate than the traditional MADM algorithm. Moreover, we also prove that our PMADM algorithm is sensitive to the parameters which vary seriously and in-sensitive to the parameters which change slightly. This property is better than that of the traditional MADM algorithm. Additionally, the PMADM algorithm is more stable than the traditional MADM algorithm. For reducing the computational complexity of the PMADM algorithm, we propose the low-complexity PMADM algorithm. For analyzing the computational complexity of the lPMADM algorithm, we propose the tree-based decomposing algorithm in this paper. The lPMADM algorithm has the same properties and performances as that of the PMADM algorithm; however, it is simpler than the PMADM algorithm. The simulation results show that the PMADM and lPMADM algorithms are much more effective than the traditional MADM algorithm.

cs.NI↗

arXiv2019

Game Theory based Joint Task Offloading and Resources Allocation Algorithm for Mobile Edge Computing

Mobile edge computing (MEC) has emerged for reducing energy consumption and latency by allowing mobile users to offload computationally intensive tasks to the MEC server. Due to the spectrum reuse in small cell network, the inter-cell interference has a great effect on MEC performances. In this paper, for reducing the energy consumption and latency of MEC, we propose a game theory based approach to join task offloading decision and resources allocation together in the MEC system. In this algorithm, the offloading decision, the CPU capacity adjustment, the transmission power control, and the network interference management of mobile users are regarded as a game. In this game, based on the best response strategy, each mobile user makes their own utility maximum rather than the utility of the whole system. We prove that this game is an exact potential game and the Nash equilibrium (NE) of this game exists. For reaching the NE, the best response approach is applied. We calculate the best response of these three variables. Moreover, we investigate the properties of this algorithm, including the convergence, the computational complexity, and the Price of anarchy (PoA). The theoretical analysis shows that the inter-cell interference affects on the performances of MEC greatly. The NE of this game is Pareto efficiency. Finally, we evaluate the performances of this algorithm by simulation. The simulation results illustrate that this algorithm is effective in improving the performances of the multi-user MEC system.

cs.NI↗

arXiv2019

Geographical and Topology Control based Opportunistic Routing for Ad Hoc Networks

The opportunistic routing has great advantages on improving packet delivery probability between the source node and candidate forwarding set. For improving and reducing energy consumption and network interference, in this paper, we propose an efficient and reliable transmission power control based opportunistic routing (ERTO) for wireless ad hoc networks. In ERTO, the network interference and fading, which are critical to routing performances but not taken into account in previous works, are considered during the prediction of. The, the expected energy consumption, and the relationship between transmission power and node degree are considered to optimize the optimal transmission power and forwarding node degree jointly. For improving routing effectiveness and reducing computational complexity, we introduce the multi-objective optimization and Pareto optimal into ERTO. During the routing process, nodes calculate the optimal transmission power and forwarding node degree according to the properties of the Pareto optimal solution set. Based on these innovations, the energy consumption, the transmission delay, and the throughout have been improved greatly.

cs.NI↗