arXiv subjects

Jun Zhuang

Publications and source records attributed to Jun Zhuang.

At least 19 recordsLinked to original sources

arXiv2026

SLT: Robust Quantum Neural Networks for Noisy-Label Medical Image Classification via Supermartingale-based Label Transition

Noisy-label learning in small-scale medical image classification is challenging and hinders the superiority of deep neural networks. Recent studies suggest that quantum neural networks (QNNs) have shown potential in limited-data regimes, yet their use for noisy-label learning remains under-explored. A key obstacle is QNNs' intrinsic "natural smoothness", which may regularize training but also obscure high-confidence samples needed for noise-transition estimation. We propose Supermartingale-based Label Transition (SLT), an anchor-free loss correction framework for robust QNN-based medical image classification under noisy labels. SLT models entropy reduction in predictive distributions as a supermartingale and uses its monotonic behavior to identify stable transition-matrix refinement steps. This enables dynamic transition updates while reducing noise-driven oscillations during QNN training. We further provide a convergence analysis showing that the proposed transition-refinement process reaches a steady state. Experiments on multiple public small-scale medical image datasets demonstrate that SLT consistently improves QNN-based classification and stably outperforms classic noise-label learning baselines under synthetic and real-world label noise.

cs.CV↗

arXiv2026

Mitigating Scaffolding Collapse in Socratic Tutors via Representation Alignment

Large language model (LLM)-based Socratic tutors increasingly guide students through multi-turn questioning, but they can suffer from scaffolding collapse: under sustained student pressure, a tutor gradually abandons guided inquiry and reveals solutions directly. Prior defenses primarily constrain observable responses through prompting, preference optimization, or filtering, leaving the internal representation drift that precedes trajectory-level collapse largely unaddressed. We propose Scaffold-Preserving Representation Alignment, a two-stage framework that first warms up a Socratic tutor with supervised fine-tuning, then combines trajectory-weighted direct preference optimization with a margin-preserving representation loss anchored to frozen reference states. Our method is designed to maintain separation between scaffold-preserving and collapse-inducing hidden states across dialogue turns. We evaluate our method across five STEM disciplines and five red-teaming attack strategies. On Qwen3-8B, our method lowers Collapse Rate to 32%, delays average collapse onset beyond nine turns, and keeps over-refusal low, suggesting that representation-level alignment can improve the robustness of long-horizon Socratic tutoring under our red-teaming protocol.

cs.AI↗

arXiv2026

PaQ-DETR: Learning Pattern and Quality-Aware Dynamic Queries for Object Detection

Detection Transformer (DETR) has redefined object detection by casting it as a set prediction task within an end-to-end framework. Despite its elegance, DETR and its variants still rely on fixed learnable queries and suffer from severe query utilization imbalance, which limits adaptability and leaves the model capacity underused. We propose PaQ-DETR (Pattern and Quality-Aware DETR), a unified framework that enhances both query adaptivity and supervision balance. It learns a compact set of shared latent patterns capturing global semantics and dynamically generates image-specific queries through content-conditioned weighting. In parallel, a quality-aware one-to-many assignment strategy adaptively selects positive samples based on localizatio-classification consistency, enriching supervision and promoting balanced query optimization. Experiments on COCO, CityScapes, and other benchmarks show consistent gains of 1.5%-4.2% mAP across DETR backbones, including ResNet and Swin-Transformer. Beyond accuracy improvement, our method provides interpretable insights into how dynamic patterns cluster semantically across object categories.

cs.CV↗

arXiv2026

Now You Hear Me: Audio Narrative Attacks Against Large Audio-Language Models

Large audio-language models increasingly operate on raw speech inputs, enabling more seamless integration across domains such as voice assistants, education, and clinical triage. This transition, however, introduces a distinct class of vulnerabilities that remain largely uncharacterized. We examine the security implications of this modality shift by designing a text-to-audio jailbreak that embeds disallowed directives within a narrative-style audio stream. The attack leverages an advanced instruction-following text-to-speech (TTS) model to exploit structural and acoustic properties, thereby circumventing safety mechanisms primarily calibrated for text. When delivered through synthetic speech, the narrative format elicits restricted outputs from state-of-the-art models, including Gemini 2.0 Flash, achieving a 98.26% success rate that substantially exceeds text-only baselines. These results highlight the need for safety frameworks that jointly reason over linguistic and paralinguistic representations, particularly as speech-based interfaces become more prevalent.

cs.CL↗

arXiv2025

Fairness-Aware Graph Representation Learning with Limited Demographic Information

Ensuring fairness in Graph Neural Networks is fundamental to promoting trustworthy and socially responsible machine learning systems. In response, numerous fair graph learning methods have been proposed in recent years. However, most of them assume full access to demographic information, a requirement rarely met in practice due to privacy, legal, or regulatory restrictions. To this end, this paper introduces a novel fair graph learning framework that mitigates bias in graph learning under limited demographic information. Specifically, we propose a mechanism guided by partial demographic data to generate proxies for demographic information and design a strategy that enforces consistent node embeddings across demographic groups. In addition, we develop an adaptive confidence strategy that dynamically adjusts each node's contribution to fairness and utility based on prediction confidence. We further provide theoretical analysis demonstrating that our framework, FairGLite, achieves provable upper bounds on group fairness metrics, offering formal guarantees for bias mitigation. Through extensive experiments on multiple datasets and fair graph learning frameworks, we demonstrate the framework's effectiveness in both mitigating bias and maintaining model utility.

cs.LG↗

arXiv2025

Learning to Conceal Risk: Controllable Multi-turn Red Teaming for LLMs in the Financial Domain

Large Language Models (LLMs) are increasingly deployed in finance, where unsafe behavior can lead to serious regulatory risks. However, most red-teaming research focuses on overtly harmful content and overlooks attacks that appear legitimate on the surface yet induce regulatory-violating responses. We address this gap by introducing a controllable black-box multi-turn risk-concealed red-teaming framework (CoRT) that progressively conceals surface-level risk while exploiting regulatory-violating behaviors. CoRT contains two key components: (i) a Risk Concealment Attacker (RCA) that generates multi-turn prompts via iterative refinement, and (ii) a Risk Concealment Controller (RCC) that predicts a turn-level Risk Concealment Score (RCS) to steer RCA's follow-up style. We also built a domain-specific benchmark, FinRisk-Bench, with 522 instructions spanning six financial risk categories. Experiments on nine widely used LLMs show that CoRT (RCA) achieves 93.19% average attack success rate (ASR), and CoRT (RCA+RCC) further improves the average ASR to 95.00%. Our code and FinRisk-Bench are available at https://github.com/gcheng128/CoRT.

cs.CL↗

arXiv2025

MediQ-GAN: Quantum-Inspired GAN for High Resolution Medical Image Generation

Machine learning-assisted diagnosis shows promise, yet medical imaging datasets are often scarce, imbalanced, and constrained by privacy, making data augmentation essential. Classical generative models typically demand extensive computational and sample resources. Quantum computing offers a promising alternative, but existing quantum-based image generation methods remain limited in scale and often face barren plateaus. We present MediQ-GAN, a quantum-inspired GAN with prototype-guided skip connections and a dual-stream generator that fuses classical and quantum-inspired branches. Its variational quantum circuits inherently preserve full-rank mappings, avoid rank collapse, and are theory-guided to balance expressivity with trainability. Beyond generation quality, we provide the first latent-geometry and rank-based analysis of quantum-inspired GANs, offering theoretical insight into their performance. Across three medical imaging datasets, MediQ-GAN outperforms state-of-the-art GANs and diffusion models. While validated on IBM hardware for robustness, our contribution is hardware-agnostic, offering a scalable and data-efficient framework for medical image generation and augmentation.

cs.CV↗

arXiv2025

InfoFlood: Jailbreaking Large Language Models with Information Overload

Large Language Models (LLMs) have demonstrated remarkable capabilities across various domains. However, their potential to generate harmful responses has raised significant societal and regulatory concerns, especially when manipulated by adversarial techniques known as "jailbreak" attacks. Existing jailbreak methods typically involve appending carefully crafted prefixes or suffixes to malicious prompts in order to bypass the built-in safety mechanisms of these models. In this work, we identify a new vulnerability in which excessive linguistic complexity can disrupt built-in safety mechanisms-without the need for any added prefixes or suffixes-allowing attackers to elicit harmful outputs directly. We refer to this phenomenon as Information Overload. To automatically exploit this vulnerability, we propose InfoFlood, a jailbreak attack that transforms malicious queries into complex, information-overloaded queries capable of bypassing built-in safety mechanisms. Specifically, InfoFlood: (1) uses linguistic transformations to rephrase malicious queries, (2) identifies the root cause of failure when an attempt is unsuccessful, and (3) refines the prompt's linguistic structure to address the failure while preserving its malicious intent. We empirically validate the effectiveness of InfoFlood on four widely used LLMs-GPT-4o, GPT-3.5-turbo, Gemini 2.0, and LLaMA 3.1-by measuring their jailbreak success rates. InfoFlood consistently outperforms baseline attacks, achieving up to 3 times higher success rates across multiple jailbreak benchmarks. Furthermore, we demonstrate that commonly adopted post-processing defenses, including OpenAI's Moderation API, Perspective API, and SmoothLLM, fail to mitigate these attacks. This highlights a critical weakness in traditional AI safety guardrails when confronted with information overload-based jailbreaks.

cs.CR↗

arXiv2025

Exploring the Vulnerability of the Content Moderation Guardrail in Large Language Models via Intent Manipulation

Intent detection, a core component of natural language understanding, has considerably evolved as a crucial mechanism in safeguarding large language models (LLMs). While prior work has applied intent detection to enhance LLMs' moderation guardrails, showing a significant success against content-level jailbreaks, the robustness of these intent-aware guardrails under malicious manipulations remains under-explored. In this work, we investigate the vulnerability of intent-aware guardrails and demonstrate that LLMs exhibit implicit intent detection capabilities. We propose a two-stage intent-based prompt-refinement framework, IntentPrompt, that first transforms harmful inquiries into structured outlines and further reframes them into declarative-style narratives by iteratively optimizing prompts via feedback loops to enhance jailbreak success for red-teaming purposes. Extensive experiments across four public benchmarks and various black-box LLMs indicate that our framework consistently outperforms several cutting-edge jailbreak methods and evades even advanced Intent Analysis (IA) and Chain-of-Thought (CoT)-based defenses. Specifically, our "FSTR+SPIN" variant achieves attack success rates ranging from 88.25% to 96.54% against CoT-based defenses on the o1 model, and from 86.75% to 97.12% on the GPT-4o model under IA-based defenses. These findings highlight a critical weakness in LLMs' safety mechanisms and suggest that intent manipulation poses a growing challenge to content moderation guardrails.

cs.CL↗

arXiv2025

Digital Forensics in the Age of Large Language Models

Digital forensics plays a pivotal role in modern investigative processes, utilizing specialized methods to systematically collect, analyze, and interpret digital evidence for judicial proceedings. However, traditional digital forensic techniques are primarily based on manual labor-intensive processes, which become increasingly insufficient with the rapid growth and complexity of digital data. To this end, Large Language Models (LLMs) have emerged as powerful tools capable of automating and enhancing various digital forensic tasks, significantly transforming the field. Despite the strides made, general practitioners and forensic experts often lack a comprehensive understanding of the capabilities, principles, and limitations of LLM, which limits the full potential of LLM in forensic applications. To fill this gap, this paper aims to provide an accessible and systematic overview of how LLM has revolutionized the digital forensics approach. Specifically, it takes a look at the basic concepts of digital forensics, as well as the evolution of LLM, and emphasizes the superior capabilities of LLM. To connect theory and practice, relevant examples and real-world scenarios are discussed. We also critically analyze the current limitations of applying LLMs to digital forensics, including issues related to illusion, interpretability, bias, and ethical considerations. In addition, this paper outlines the prospects for future research, highlighting the need for effective use of LLMs for transparency, accountability, and robust standardization in the forensic process.

cs.CR↗

arXiv2025

Large Language Models Can Help Mitigate Barren Plateaus in Quantum Neural Networks

In the era of noisy intermediate-scale quantum (NISQ) computing, Quantum Neural Networks (QNNs) have emerged as a promising approach for various applications, yet their training is often hindered by barren plateaus (BPs), where gradient variance vanishes exponentially as the qubit size increases. Most initialization-based mitigation strategies rely heavily on pre-designed static parameter distributions, thereby lacking adaptability to diverse model sizes or data conditions. To address these limitations, we propose AdaInit, a foundational framework that leverages large language models with the submartingale property to iteratively synthesize initial parameters for QNNs that yield non-negligible gradient variance, thereby mitigating BPs. Unlike conventional one-shot initialization methods, AdaInit adaptively explores the parameter space by incorporating dataset characteristics and gradient feedback, with theoretical guarantees of convergence to finding a set of effective initial parameters for QNNs. We provide rigorous theoretical analyses of the submartingale-based process and empirically validate that AdaInit consistently outperforms existing initialization methods in maintaining higher gradient variance across various QNN scales. We believe this work may initiate a new avenue to mitigate BPs.

quant-ph↗

arXiv2025

Investigating and Mitigating Barren Plateaus in Variational Quantum Circuits: A Survey

In recent years, variational quantum circuits (VQCs) have been widely explored to advance quantum circuits against classic models on various domains, such as quantum chemistry and quantum machine learning. Similar to classic machine-learning models, VQCs can be trained through various optimization approaches, such as gradient-based or gradient-free methods. However, when employing gradient-based methods, the gradient variance of VQCs may dramatically vanish as the number of qubits or layers increases. This issue, a.k.a. Barren Plateaus (BPs), seriously hinders the scaling of VQCs on large datasets. To mitigate the barren plateaus, extensive efforts have been devoted to tackling this issue through diverse strategies. In this survey, we conduct a systematic literature review of recent works from both investigation and mitigation perspectives. Furthermore, we propose a new taxonomy to categorize most existing mitigation strategies into five groups and introduce them in detail. Also, we compare the concurrent survey papers about BPs. Finally, we provide insightful discussion on future directions for BPs.

quant-ph↗

arXiv2025

Enhanced Atom-by-Atom Assembly of Defect-Free Two-Dimensional Mixed-Species Atomic Arrays

Defect-free single atom array in optical tweezers is a promising platform for scalable quantum computing, quantum simulation, and quantum metrology. Extending single-species array to mixed-species one promise to offer new possibilities. In our recent proof of principle realization of defect-free two-dimensional assembly of mixed-species $^{85}$Rb ($^{87}$Rb) atom arrays [C. Sheng et al.\href{https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.128.083202}{{\color{blue} Phys. Rev. Lett. 128, 083202(2022)}}], the filling fractions were limited by the imperfect transfer of atoms and the occurrence of logjams during the atom rearrangement. In order to scale up the size of defect-free mixed-species atom array, we scale up the tweezer array and improve the atom transfer, and upgrade the heuristic heteronuclear algorithm so as to facilitate multiple rearrangement cycles. Consequently, we successfully create defect-free atom arrays with 120 mixed-species single atoms. The corresponding filling fraction and defect-free probability are improved to be 98.6(1)\% and 14(2)\%, respectively. It is anticipated that the enhanced algorithm can be extended to other combinations of atomic species, and this mixed-species atom array is readily for studies of many-body physics, quantum error correction, and quantum metrology.

quant-ph↗

arXiv2024

Blockchain for Large Language Model Security and Safety: A Holistic Survey

With the growing development and deployment of large language models (LLMs) in both industrial and academic fields, their security and safety concerns have become increasingly critical. However, recent studies indicate that LLMs face numerous vulnerabilities, including data poisoning, prompt injections, and unauthorized data exposure, which conventional methods have struggled to address fully. In parallel, blockchain technology, known for its data immutability and decentralized structure, offers a promising foundation for safeguarding LLMs. In this survey, we aim to comprehensively assess how to leverage blockchain technology to enhance LLMs' security and safety. Besides, we propose a new taxonomy of blockchain for large language models (BC4LLMs) to systematically categorize related works in this emerging field. Our analysis includes novel frameworks and definitions to delineate security and safety in the context of BC4LLMs, highlighting potential research directions and challenges at this intersection. Through this study, we aim to stimulate targeted advancements in blockchain-integrated LLM security.

cs.CR↗

arXiv2024

Deconstructing The Ethics of Large Language Models from Long-standing Issues to New-emerging Dilemmas: A Survey

Large Language Models (LLMs) have achieved unparalleled success across diverse language modeling tasks in recent years. However, this progress has also intensified ethical concerns, impacting the deployment of LLMs in everyday contexts. This paper provides a comprehensive survey of ethical challenges associated with LLMs, from longstanding issues such as copyright infringement, systematic bias, and data privacy, to emerging problems like truthfulness and social norms. We critically analyze existing research aimed at understanding, examining, and mitigating these ethical risks. Our survey underscores integrating ethical standards and societal values into the development of LLMs, thereby guiding the development of responsible and ethically aligned language models.

cs.CL↗

arXiv2024

A Survey on the Application of Generative Adversarial Networks in Cybersecurity: Prospective, Direction and Open Research Scopes

With the proliferation of Artificial Intelligence, there has been a massive increase in the amount of data required to be accumulated and disseminated digitally. As the data are available online in digital landscapes with complex and sophisticated infrastructures, it is crucial to implement various defense mechanisms based on cybersecurity. Generative Adversarial Networks (GANs), which are deep learning models, have emerged as powerful solutions for addressing the constantly changing security issues. This survey studies the significance of the deep learning model, precisely on GANs, in strengthening cybersecurity defenses. Our survey aims to explore the various works completed in GANs, such as Intrusion Detection Systems (IDS), Mobile and Network Trespass, BotNet Detection, and Malware Detection. The focus is to examine how GANs can be influential tools to strengthen cybersecurity defenses in these domains. Further, the paper discusses the challenges and constraints of using GANs in these areas and suggests future research directions. Overall, the paper highlights the potential of GANs in enhancing cybersecurity measures and addresses the need for further exploration in this field.

cs.CR↗

arXiv2024

RW-NSGCN: A Robust Approach to Structural Attacks via Negative Sampling

Node classification using Graph Neural Networks (GNNs) has been widely applied in various practical scenarios, such as predicting user interests and detecting communities in social networks. However, recent studies have shown that graph-structured networks often contain potential noise and attacks, in the form of topological perturbations and weight disturbances, which can lead to decreased classification performance in GNNs. To improve the robustness of the model, we propose a novel method: Random Walk Negative Sampling Graph Convolutional Network (RW-NSGCN). Specifically, RW-NSGCN integrates the Random Walk with Restart (RWR) and PageRank (PGR) algorithms for negative sampling and employs a Determinantal Point Process (DPP)-based GCN for convolution operations. RWR leverages both global and local information to manage noise and local variations, while PGR assesses node importance to stabilize the topological structure. The DPP-based GCN ensures diversity among negative samples and aggregates their features to produce robust node embeddings, thereby improving classification performance. Experimental results demonstrate that the RW-NSGCN model effectively addresses network topology attacks and weight instability, increasing the accuracy of anomaly detection and overall stability. In terms of classification accuracy, RW-NSGCN significantly outperforms existing methods, showing greater resilience across various scenarios and effectively mitigating the impact of such vulnerabilities.

cs.LG↗

arXiv2024

Probing the interaction energy of two $^{85}$Rb atoms in an optical tweezer via spin-motion coupling

The inherent polarization gradients in tight optical tweezers can be used to couple the atomic spins to the two-body motion under the action of a microwave spin-flip transition, so that such a spin-motion coupling offers an important control knob on the motional states of optically trapped two colliding atoms. Here, after preparing two elastically scattering $^{85}$Rb atoms in the three-dimensional ground-state in the optical tweezer, we employed this control in order to probe the colliding energies of elastic and inelastic channels. The combination of microwave spectra and corresponding s-wave pseudopotential model allows us to infer the effect of the state-dependent trapping potentials on the elastic colliding energies, as well as to reveal how the presence of inelastic interactions affects elastic part of the relative potential. Our work shows that the spin-motion coupling in a tight optical tweezer expand the experimental toolbox for fundamental studies of ultracold collisions in the two body systems with reactive collisions, and potentially for that of more complex interactions, such as optically trapped atom-molecule and molecule-molecule interactions.

physics.atom-ph↗