Searcharxiv⌕ Search

arXiv subjects

Shuning Zhang

Publications and source records attributed to Shuning Zhang.

At least 19 recordsLinked to original sources
arXiv2026

Air-Ground Collaborative Vision-and-Language Navigation via Shared Bird's-Eye Maps

Air-ground collaborative Vision-and-Language Navigation (VLN) pairs an unmanned aerial vehicle (UAV) with a global bird's-eye view and an unmanned ground vehicle (UGV) with a local first-person view, yet the setting remains largely unexplored: existing training-free methods solve single-agent tasks but offer no collaboration mechanism, and a recent CARLA-Air evaluation found no stable cooperative behavior across five state-of-the-art VLA models; naive semantic communication or bidirectional coupling even degrades performance. We establish AGC-VLN (Air-Ground Collaborative VLN), the first training-free baseline for air-ground collaborative VLN. The key insight is that training-free methods decompose navigation into VLM-based semantic reasoning and deterministic geometric execution, exposing a collaboration interface: the UAV's global view, over which it renders the UGV's reported pose and the VLM-anchored target as CAR/GOAL markers with distance labels, yielding a shared bird's-eye map. From this map, the UGV acquires global spatial context its first-person view cannot provide, plans a road-following path with a frozen VLM, and executes it under closed-loop control; in parallel, the UAV runs 3D-SPF, a spatial-search upgrade of SPF that localizes the target in the downward view and flies toward it. On 100 closed-loop episodes in CARLA-Air's Town10HD scene, AGC-VLN reaches a 77.0% joint success rate, a collaboration gain of +27.0% over the weaker individual agent (the UAV, 50.0%), and exceeds the strongest published single-agent baseline (Travel UAV, 53.0%) by 24.0 points, stemming from the complementarity of the UAV's global view and the UGV's road-following execution. Project page: https://github.com/ZSN2024/AGC-VLN.

cs.RO↗
arXiv2026

CoVer: Conflict-Aware Claim Verification

Social media fact-checking has long been challenged by evidence-level and aggregation-level conflicts, where erroneous evidence mimics authoritative news sources. To capture this challenge and support conflict verification tasks, we present ContraNote, a large-scale real-world dataset curated from X's Community Notes system. It includes 33,686 posts for evaluating evidence-level conflict resolution, and 54,474 instances for evaluating aggregation-level prioritization. Additionally, we propose CoVer, a factual adjudication framework with three-stage pipelines: evidence schema normalization, factual consensus and support verification. This prioritizes evidence over noise to prevent it from compromising the final verdict. Technical evaluations show that CoVer achieves strong performance compared with state-of-the-art baselines across ContraNote (86.0% Acc., 68.0% mac. F1, 64.5 bal. Acc. on Conflict; and 88.5% Acc., 88.5 mac. F1 and 89.2 bal. Acc. on Prioritization), CONFACT-HumC (88.4% Acc.) and CONFACT-ModC (89.4% Acc.).

cs.AI↗
arXiv2026

Position: Robot Privacy as Embodied Boundary Work. Connecting Capabilities, Contexts, and Design Responses in Everyday Robotics

Robots are increasingly entering everyday environments where privacy is shaped not only by data practices, but also by spatial, bodily, social, and relational boundaries. Their embodied capabilities allow them to reshape these boundaries through situated action, challenging privacy framings centered on data flows, interface settings, or one-time consent. Prior work has examined robot privacy through sensing, data collection, telepresence, transparency, consent, bystander awareness, and multi-stakeholder governance. Building on this work, we propose embodied boundary privacy as a capability-by-context framing for examining how physically present robots may reshape privacy boundaries in situated interaction. Specifically, this framing organizes privacy risks across seven robot capabilities and five deployment contexts, asking how embodied capabilities enable boundary crossings and how situated contexts shape who is affected, how these crossings are interpreted, and when they become contested. We use this perspective to outline design and research implications for embodied privacy mechanisms, including boundary checkpoints, viewpoint-aware sensing control, remote-presence disclosure, object- and body-level access rules, constraints on socially persuasive privacy influence, and local interruption rights. We encourage HRI research, design, and governance to treat robot movement, orientation, proximity, object access, remote presence, and social expression as privacy-relevant actions whose meaning depends on context.

cs.RO↗
arXiv2026

Understanding Security and Privacy Perceptions of Content Creators Regarding AI Labels of AI-Generated Content

AI labels, typically implemented via underlying tracing mechanisms such as watermarks and metadata, are crucial for protecting Artificial Intelligence-Generated Content (AIGC) against security threats like disinformation and evasion. However, the perceived devaluation of AI-assisted work discourages creators from disclosing AI use, incentivizing efforts to bypass labeling and compromising downstream traceability. Yet, how AIGC creators perceive the security and privacy (S\&P) implications of these labels, and how their behaviors impact technical resilience remain underexplored. To this end, we conducted semi-structured interviews with 21 AIGC creators and measured images across 6 image generation platforms against 16 self-reported manipulation settings. Our findings reveal that creators conflate binary AI labels with granular traceability, and express strong fears of de-anonymization via platform identifiers. Driven by fears of algorithmic traffic suppression and reputational risks, they defensively removed digital traces. Through empirical tests, we show that targeted modifications like coarse quantization significantly degrade detection. AI detection capabilities are also inconsistent across platforms, and suffer from false positives even for human-authored images. Based on these insights, we advocate for workflow-resilient implicit AI labels that align technical guarantees with creators' incentives.

cs.HC↗
arXiv2026

Covert Semantic Transmission in ISAC: Dual-Functional Waveform Design and Rectified Flow-Assisted Recovery

Semantic integrated sensing and communication (ISAC) is envisioned as a promising paradigm for efficient and intelligent connectivity in future wireless networks. However, the open wireless channel exposes the dual-functional waveform to detection, which challenges the joint guarantee of covertness, sensing fidelity, and semantic accuracy. To address the challenge, we propose CoSMIC, a novel covertness-oriented semantic ISAC framework, where the sensing output is embedded into a dual-functional ISAC waveform through semantic modulation. Specifically, a semantic rotation coding scheme is established to map semantic latents onto the pairwise rotation and scaling of Gaussian reference sequences, which satisfies a derived closed-form covertness constraint by a differentiable budget projection. Moreover, the radar performance is analyzed to confirm an invariant matched-filter mainlobe response and a bounded output signal-to-interference-plus-noise ratio (SINR) under the semantic embedding. Subsequently, a reliability-guided rectified flow (RFlow) refiner is designed to effectively reconstruct high-fidelity semantic representations from coarse observations. Simulation results demonstrate that CoSMIC improves the semantic reconstruction quality by 18% over diffusion-based baseline schemes with substantially reduced inference latency under strict covertness constraints, which validates the applicability to practical ISAC scenarios. The source code and video demonstrations are available at https://github.com/LanceAnlan/CoSMIC-covertness-oriented-semantic-ISAC-framework.

eess.SP↗
arXiv2026

CANote: Empowering Fact-checking Note Writing Through Scaffolded and Provenance-based Human-AI Collaboration

Crowdsourced fact-checking mechanisms, such as X's Community Notes, play a critical role in mitigating the spread of misinformation. However, drafting high-quality, evidence-based debunking notes imposes a substantial burden on contributors. We present CANote, an AI-assisted debunking note writing system featuring evidence correlation and structured co-drafting. CANote scaffolds the workflow by extracting subclaims from social media posts, providing provenance through explicit links between subclaims and retrieved evidence, and generating neutral, structural drafts to support human reasoning. We evaluated CANote against manual writing (N=52 fact-checkers, N=52 lay users) on simulated X platform, where we found CANote significantly improves note quality. Notably, CANote enables lay users to write notes that have comparable quality to those written by experts. While the task completion time and perceived cognitive load remain comparable to manual drafting, CANote significantly increases user satisfaction. However, this assistance introduces a trade-off, resulting in a reduced sense of user ownership and control over the debunking note.

cs.HC↗
arXiv2026

Towards Multi-Agent-Simulation-Based Community Note Evaluation

Community-based fact-checking that relies on cross-consensus is expanding rapidly on social media platforms. However, the delay and low-ratio of cross-consensus community fact-checks rated by human contributors remains a significant challenge. To address this, we first created ComRate, a large-scale dataset comprising 2.5 million community notes and over 209 million ratings sourced from $\mathbb{X}$. We then propose MultiCom, a persona-guided multi-agent rating framework for community note evaluation. MultiCom simulates diverse rater population by clustering contributors in a matrix-factorized rater space and prompting persona agents to generate structured assessments based on the official community notes rating schema. These agents output structured and explainable judgments, such as confidence, agreement signals and reasons. An out-of-fold calibrated aggregation algorithm combines features such as raw votes and diagnostic reason signals for reliable prediction. Extensive evaluations demonstrate that MultiCom outperforms alternative methods, achieving an average accuracy of 84.7% (balanced accuracy 68.3%, macro-F1 60.1%) on the evaluation set.

cs.SI↗
arXiv2026

Focused on the User, Overlooking the Risks: Security and Privacy Understandings, Practices and Challenges of Independent Chinese AI Agent Developers

The proliferation of AI agents empowers independent developers, defined as individual or small groups who self-initiate projects rather than fulfill client-based contracts, to create sophisticated autonomous systems, but also introduces novel security and privacy (S&P) challenges beyond traditional corporate structures. We conducted an interview study (N=28) with Chinese developers, whose extensive use of global LLM services offer valuable insights into this population. We investigate their understandings, practices and challenges of S&P challenges in their developed AI agent products. We revealed that independent developers frequently think and act from their users' perspective. They focused on user-facing safety risks such as harmful content while exhibiting low awareness of security vulnerabilities. Consequently, developers rely almost exclusively on ad-hoc, manually crafted safeguards and informal communication, with an absence of formal tools or processes for S&P practices. We found these actions are driven by various inhibitors, primarily a lack of formal training on S&P related skills, accessible security tools and actionable guidance from platforms. Our work contributed the first exploration of independent AI agent developers' S&P understanding, outlining opportunities for tailored security tooling.

cs.HC↗
arXiv2026

Generative AI-Enabled Refund Fraud in Chinese E-Commerce: Investigation on Merchants and Platform Workers

E-commerce dispute resolution typically relies on the security assumption that digital evidence truthfully reflects physical reality. Generative AI (GenAI) invalidates this threat model, enabling attackers to fabricate hyper-realistic evidence of product defects at negligible cost. Through semi-structured interviews with merchants (N=17) and platform workers (N=13) in the Chinese e-commerce market, we characterize this shift toward GenAI-enabled scalable fabrication. We outline a taxonomy of four GenAI-enabled threat vectors across the transaction, dispute, logistics and communication phases, highlighting how attackers exploit GenAI to synthesize physically plausible product defects at scale. To mitigate these threats, platforms and merchants are adapting verification strategies, relying on AI tools for automated screening and adversarial interrogation (e.g., requesting multi-angle videos) to increase attack complexity. However, we find several challenges that hinder the adoption of these defenses, including implementation hurdles like structural platform constraints and fundamental limitations regarding the technical sophistication of GenAI. We conclude by outlining design implications for privacy-preserving cross-platform fraud databases, and traceability mechanisms such as embedding verifiable material anchors into the product.

cs.CR↗
arXiv2026

Investigating Novice Researchers' Perceptions of Research Privacy Within LLM-Assisted Workflows

Large Language Model (LLMs)-assisted scholarly workflows introduce critical privacy and intellectual property risks. As a uniquely vulnerable cohort driven by publication pressure and a lack of institutional support, novice researchers rely heavily on public LLMs, compelling them to navigate high-stakes privacy-publication trade-offs. To investigate these concerns, we conducted semi-structured interviews with 44 researchers across diverse disciplines. Our findings reveal that the fear of idea leakage paradoxically accelerates, rather than deters, reliance on LLMs, as researchers utilize them to expedite publication. They also held misconceptions that their ideas lacked the unique value to attract targeted attacks, and that their inputs would be safely diluted within massive datasets, preventing reconstruction. From interviews, we identified five types of mitigations including input fragmentation and adversarial probing, though we found that participants largely perceived these measures as ineffective. We outline implications including implementing institution-level sandboxed isolation, scenario-based privacy pedagogy, and verifiable data-deletion audits for transparency.

cs.HC↗
arXiv2026

UniGround: Universal 3D Visual Grounding via Training-Free Scene Parsing

3D Visual Grounding (3DVG) localizes objects from natural-language descriptions in 3D scenes and is fundamental to embodied AI applications. Although foundation models enable open-vocabulary reasoning, they typically rely on pre-generated candidates, creating two sequential bottlenecks. The \emph{candidate bottleneck} occurs when dataset-specific 3D proposal models miss, fragment, or incorrectly group targets under distribution shifts, excluding them from VLM reasoning. The \emph{evidence bottleneck} stems from incomplete visual evidence: global renderings preserve spatial context but obscure object details, whereas candidate-centric views capture local appearance but lack global context. To address these bottlenecks, we propose UniGround, a zero-shot 3DVG framework that addresses both bottlenecks through Global Candidate Filtering and Contextual Precision Grounding. Global Candidate Filtering constructs topology-consistent, class-agnostic candidates from 3D topology and multi-view semantic cues, without dataset-trained 3D detectors, task-specific proposal supervision, or predefined box and category priors. Contextual Precision Grounding jointly reasons over global spatial context and candidate-centric visual evidence, followed by closed-loop consistency verification for reliable target identification. UniGround achieves 46.1\%/34.1\% Acc@0.25/0.5 on ScanRefer and 28.7\% Acc@0.25 on the evaluated ARKitScenes subset of EmbodiedScan. Further experiments demonstrate competitive grounding without dataset-specific 3D priors, cross-dataset generalization to unseen indoor scenes, and robustness to real-world reconstruction noise and practical domain shifts.

cs.RO↗
arXiv2026

ModalImmune: Immunity Driven Unlearning via Self Destructive Training

Multimodal systems are vulnerable to partial or complete loss of input channels at deployment, which undermines reliability in real-world settings. This paper presents ModalImmune, a training framework that enforces modality immunity by intentionally and controllably collapsing selected modality information during training so the model learns joint representations that are robust to destructive modality influence. The framework combines a spectrum-adaptive collapse regularizer, an information-gain guided controller for targeted interventions, curvature-aware gradient masking to stabilize destructive updates, and a certified Neumann-truncated hyper-gradient procedure for automatic meta-parameter adaptation. Empirical evaluation on standard multimodal benchmarks demonstrates that ModalImmune improves resilience to modality removal and corruption while retaining convergence stability and reconstruction capacity.

cs.LG↗
arXiv2026

zkCraft: Prompt-Guided LLM as a Zero-Shot Mutation Pattern Oracle for TCCT-Powered ZK Fuzzing

Zero-knowledge circuits enable privacy-preserving and scalable systems but are difficult to implement correctly due to the tight coupling between witness computation and circuit constraints. We present zkCraft, a practical framework that combines deterministic, R1CS-aware localization with proof-bearing search to detect semantic inconsistencies. zkCraft encodes candidate constraint edits into a single Row-Vortex polynomial and replaces repeated solver queries with a Violation IOP that certifies the existence of edits together with a succinct proof. Deterministic LLM-driven mutation templates bias exploration toward edge cases while preserving auditable algebraic verification. Evaluation on real Circom code shows that proof-bearing localization detects diverse under- and over-constrained faults with low false positives and reduces costly solver interaction. Our approach bridges formal verification and automated debugging, offering a scalable path for robust ZK circuit development.

cs.CR↗
arXiv2026

VisGuardian: A Lightweight Group-based Privacy Control Technique For Front Camera Data From AR Glasses in Home Environments

Always-on sensing of AI applications on AR glasses makes traditional permission techniques ill-suited for context-dependent visual data, especially within home environments. The home presents a highly challenging privacy context due to the high density of sensitive objects, and the frequent presence of non-consenting family members, and the intimate nature of daily routines, making it a critical focus area for scalable privacy control mechanisms. Existing fine-grained controls, while offering nuanced choices, are inefficient for managing multiple private objects. We propose VisGuardian, a fine-grained content-based visual permission technique for AR glasses. VisGuardian features a group-based control mechanism that enables users to efficiently manage permissions for multiple private objects. VisGuardian detects objects using YOLO and adopts a pre-classified schema to group them. By selecting a single object, users can efficiently obscure groups of related objects based on criteria including privacy sensitivity, object category, or spatial proximity. A technical evaluation shows VisGuardian achieves mAP50 of 0.6704 with only 14.0 ms latency and a 1.7% increase in battery consumption per hour. Furthermore, a user study (N=24) comparing VisGuardian to slider-based and object-based baselines found it to be significantly faster for setting permissions and was preferred by users for its efficiency, effectiveness, and ease of use.

cs.HC↗
arXiv2026

A Scoping Review and Guidelines on Privacy Policy's Visualization from an HCI Perspective

Privacy Policies are a cornerstone of informed consent, yet a persistent gap exists between their legal intent and practical efficacy. Despite decades of Human-Computer Interaction (HCI) research proposing various visualizations, user comprehension remains low, and designs rarely see widespread adoption. To understand this landscape and chart a path forward, we synthesized 65 top-tier papers using a framework adapted from the user-centered design lifecycle. Our analysis presented findings of the field's evolution across four dimensions: (1) the trade-off between information load and decision efficacy, which demonstrates a shift from augmenting disclosures to prioritizing information condensation and cognitive load management to counter the inefficacy of comprehensive texts, (2) the co-evolutionary dynamic of design and automation, revealing that complex design ambitions such as context-awareness drove the need for advanced NLP, while recent LLM breakthroughs are enabling the semantic interpretation required to realize those designs, (3) the tension between generality and specificity, highlighting the divergence between standardized, cross-platform solutions and the increasing necessity for specialized, context-aware interaction patterns in IoT and immersive environments, and (4) balancing stakeholder opinions, which shows that visualization efficacy is constrained by the complex interplay of regulatory mandates, developer capabilities and provider incentives. We conclude by outlining four critical challenges for future research.

cs.HC↗
arXiv2026

Collab: Fostering Critical Identification of Deepfake Videos on Social Media via Synergistic Annotation

Identifying deepfake videos on social media platforms is challenged by dynamic spatio-temporal artifacts and inadequate user tools. This hinders both critical viewing by users and scalable moderation on platforms. Here, we present Collab, a web plugin enabling users to collaboratively annotate deepfake videos. Collab integrates three key components: (i) an intuitive interface for spatio-temporal labeling where users provide confidence scores and rationales, facilitating detailed input even from non-experts, (ii) a novel confidence-weighted spatio-temporal Intersection-over-Union (IoU) algorithm to aggregate diverse user annotations into accurate aggregations, and (iii) a hierarchical demonstration strategy presenting aggregated results to guide attention toward contentious regions and foster critical evaluation. A seven-day online study (N=90), where participants annotated suspicious videos when viewing an online experimental platforms, compared Collab against two conditions without aggregation or demonstration respectively. Collab significantly improved identification accuracy and enhanced reflection compared to non-demonstration condition, while outperforming non-aggregation condition for its novelty and effectiveness.

cs.HC↗
arXiv2026

"Privacy across the boundary": Examining Perceived Privacy Risk Across Data Transmission and Sharing Ranges of Smart Home Personal Assistants

As Smart Home Personal Assistants (SPAs) evolve into social agents, understanding user privacy necessitates interpersonal communication frameworks, such as Privacy Boundary Theory (PBT). To ground our investigation, our three-phase preliminary study (1) identified transmission and sharing ranges as key boundary-related risk factors, (2) categorized relevant SPA functions and data types, and (3) analyzed commercial practices, revealing widespread data sharing and non-transparent safeguards. A subsequent mixed-methods study (N=412 survey, N=40 interviews among the survey participants) assessed users' perceived privacy risks across data types, transmission ranges and sharing ranges. Results demonstrate a significant, non-linear escalation in perceived risk when data crosses two critical boundaries: the `public network' (transmission) and `third parties' (sharing). This boundary effect holds robustly across data types and demographics. Furthermore, risk perception is modulated by data attributes (e.g., social relational data), and contextual privacy calculus. Conversely, anonymization safeguards show limited efficacy especially for third-party sharing, a finding attributed to user distrust. These findings empirically ground PBT in the SPA context and inform design of boundary-aware privacy protection.

cs.HC↗
arXiv2025

HERO: Hierarchical Traversable 3D Scene Graphs for Embodied Navigation Among Movable Obstacles

3D Scene Graphs (3DSGs) constitute a powerful representation of the physical world, distinguished by their abilities to explicitly model the complex spatial, semantic, and functional relationships between entities, rendering a foundational understanding that enables agents to interact intelligently with their environment and execute versatile behaviors. Embodied navigation, as a crucial component of such capabilities, leverages the compact and expressive nature of 3DSGs to enable long-horizon reasoning and planning in complex, large-scale environments. However, prior works rely on a static-world assumption, defining traversable space solely based on static spatial layouts and thereby treating interactable obstacles as non-traversable. This fundamental limitation severely undermines their effectiveness in real-world scenarios, leading to limited reachability, low efficiency, and inferior extensibility. To address these issues, we propose HERO, a novel framework for constructing Hierarchical Traversable 3DSGs, that redefines traversability by modeling operable obstacles as pathways, capturing their physical interactivity, functional semantics, and the scene's relational hierarchy. The results show that, relative to its baseline, HERO reduces PL by 35.1% in partially obstructed environments and increases SR by 79.4% in fully obstructed ones, demonstrating substantially higher efficiency and reachability.

cs.RO↗