arXiv subjects

Zhaohui Chen

Publications and source records attributed to Zhaohui Chen.

13 recordsLinked to original sources

arXiv2026

More Granular, Less Trust: Enforcing Intra-Process Isolation with Arm CCA in an Untrusted Management Environment

With the increasing adoption of confidential computing, security-sensitive applications are often deployed in confidential virtual machines (CVMs), which reduce reliance on third-party cloud providers. However, privilege attacks originating from the OS remain a significant threat in these environments. Existing finer-grained isolation schemes, such as SHELTER, provide process-level protection but are still vulnerable to intraprocess attacks and potential collusion between the OS and intra-process adversaries. Many current intra-process isolation techniques continue to depend on the OS to manage and enforce isolation domains, leading to a large Trusted Computing Base (TCB). This gap highlights the need for more granular, less trust-dependent confidential computing solutions. In this paper, we present CCAegis, a system that extends the Arm Confidential Compute Architecture (CCA) to enforce intra-process isolation of sensitive data and operations, safeguarding them from both intraprocess adversaries and the OS. We employ static analysis to track the flow of sensitive data and identify functions that handle such data. Permission-switching instructions are inserted at the function call and return points, adjusting permissions via the Granule Protection Table (GPT) to ensure that only designated functions can access the isolated data. Notably, CCAegis places trust solely in the Secure Monitor, which configures the GPTs and manages domain switching, thereby minimizing the TCB. We implemented CCAegis on both an official emulator and a real development board to assess its performance. Our experimental results show that CCAegis effectively isolates sensitive data and operations, with performance overheads ranging from 1.01x to 1.43x compared to the original version across real-world cryptographic workloads.

cs.CR↗

arXiv2025

Automating construction safety inspections using a multi-modal vision-language RAG framework

Conventional construction safety inspection methods are often inefficient as they require navigating through large volume of information. Recent advances in large vision-language models (LVLMs) provide opportunities to automate safety inspections through enhanced visual and linguistic understanding. However, existing applications face limitations including irrelevant or unspecific responses, restricted modal inputs and hallucinations. Utilisation of Large Language Models (LLMs) for this purpose is constrained by availability of training data and frequently lack real-time adaptability. This study introduces SiteShield, a multi-modal LVLM-based Retrieval-Augmented Generation (RAG) framework for automating construction safety inspection reports by integrating visual and audio inputs. Using real-world data, SiteShield outperformed unimodal LLMs without RAG with an F1 score of 0.82, hamming loss of 0.04, precision of 0.76, and recall of 0.96. The findings indicate that SiteShield offers a novel pathway to enhance information retrieval and efficiency in generating safety reports.

cs.CV↗

arXiv2025

LVLMs as inspectors: an agentic framework for category-level structural defect annotation

Automated structural defect annotation is essential for ensuring infrastructure safety while minimizing the high costs and inefficiencies of manual labeling. A novel agentic annotation framework, Agent-based Defect Pattern Tagger (ADPT), is introduced that integrates Large Vision-Language Models (LVLMs) with a semantic pattern matching module and an iterative self-questioning refinement mechanism. By leveraging optimized domain-specific prompting and a recursive verification process, ADPT transforms raw visual data into high-quality, semantically labeled defect datasets without any manual supervision. Experimental results demonstrate that ADPT achieves up to 98% accuracy in distinguishing defective from non-defective images, and 85%-98% annotation accuracy across four defect categories under class-balanced settings, with 80%-92% accuracy on class-imbalanced datasets. The framework offers a scalable and cost-effective solution for high-fidelity dataset construction, providing strong support for downstream tasks such as transfer learning and domain adaptation in structural damage assessment.

cs.CV↗

arXiv2025

Entropy Measures for Transition Matrices in Random Systems

A transition matrix can be constructed through the partial contraction of two given quantum states. We analyze and compare four different definitions of entropy for transition matrices, including (modified) pseudo entropy, SVD entropy, and ABB entropy. We examine the probabilistic interpretation of each entropy measure and show that only the distillation interpretation of ABB entropy corresponds to the joint success probability of distilling entanglement between the two quantum states used to construct the transition matrix. Combining the transition matrix with preceding measurements and subsequent non-unitary operations, the ABB entropy either decreases or remains unchanged, whereas the pseudo-entropy and SVD entropy may increase or decrease. We further apply these entropy measures to transition matrices constructed from several ensembles: (i) pairs of independent Haar-random states; (ii) bi-orthogonal eigenstates of non-Hermitian random systems; and (iii) bi-orthogonal states in $PT$-symmetric systems near their exceptional points. Across all cases considered, the SVD and ABB entropies of the transition matrix closely mirror the behavior of the subsystem entanglement entropy of a single random state, in contrast to the (modified) pseudo entropy, which can exceed the bound of subsystem size, fail to scale with system size, or even take complex values.

hep-th↗

arXiv2025

Ironman: Accelerating Oblivious Transfer Extension for Privacy-Preserving AI with Near-Memory Processing

With the wide application of machine learning (ML), privacy concerns arise with user data as they may contain sensitive information. Privacy-preserving ML (PPML) based on cryptographic primitives has emerged as a promising solution in which an ML model is directly computed on the encrypted data to provide a formal privacy guarantee. However, PPML frameworks heavily rely on the oblivious transfer (OT) primitive to compute nonlinear functions. OT mainly involves the computation of single-point correlated OT (SPCOT) and learning parity with noise (LPN) operations. As OT is still computed extensively on general-purpose CPUs, it becomes the latency bottleneck of modern PPML frameworks. In this paper, we propose a novel OT accelerator, dubbed Ironman, to significantly increase the efficiency of OT and the overall PPML framework. We observe that SPCOT is computation-bounded, and thus propose a hardware-friendly SPCOT algorithm with a customized accelerator to improve SPCOT computation throughput. In contrast, LPN is memory-bandwidth-bounded due to irregular memory access patterns. Hence, we further leverage the near-memory processing (NMP) architecture equipped with memory-side cache and index sorting to improve effective memory bandwidth. With extensive experiments, we demonstrate Ironman achieves a 39.2-237.4 times improvement in OT throughput across different NMP configurations compared to the full-thread CPU implementation. For different PPML frameworks, Ironman demonstrates a 2.1-3.4 times reduction in end-to-end latency for both CNN and Transformer models.

cs.AR↗

arXiv2024

Integration of Large Vision Language Models for Efficient Post-disaster Damage Assessment and Reporting

Traditional natural disaster response involves significant coordinated teamwork where speed and efficiency are key. Nonetheless, human limitations can delay critical actions and inadvertently increase human and economic losses. Agentic Large Vision Language Models (LVLMs) offer a new avenue to address this challenge, with the potential for substantial socio-economic impact, particularly by improving resilience and resource access in underdeveloped regions. We introduce DisasTeller, the first multi-LVLM-powered framework designed to automate tasks in post-disaster management, including on-site assessment, emergency alerts, resource allocation, and recovery planning. By coordinating four specialised LVLM agents with GPT-4 as the core model, DisasTeller autonomously implements disaster response activities, reducing human execution time and optimising resource distribution. Our evaluations through both LVLMs and humans demonstrate DisasTeller's effectiveness in streamlining disaster response. This framework not only supports expert teams but also simplifies access to disaster management processes for non-experts, bridging the gap between traditional response methods and LVLM-driven efficiency.

cs.MA↗

arXiv2024

Security Issuance, Institutional Investors and Quid Pro Quo

Securities issuance through intermediaries is subject to agency problems and informational frictions. We examine these effects using SPAC data. We identify ``premium'' investors whose participation is linked to lower liquidation risk, higher returns, and lower redemption rates, consistent with both informational rents and agency frictions. In contrast, ``non-premium'' investors engage in non-agency quid pro quo relationships. Specifically, they receive high returns from an intermediary (quid) in exchange for a tacit agreement to participate in weaker future deals (quo). These relationships serve as insurance for issuers and intermediaries, enabling more issuers to access markets.

q-fin.GN↗

arXiv2024

Vision Mamba-based autonomous crack segmentation on concrete, asphalt, and masonry surfaces

Convolutional neural networks (CNNs) and Transformers have shown advanced accuracy in crack detection under certain conditions. Yet, the fixed local attention can compromise the generalisation of CNNs, and the quadratic complexity of the global self-attention restricts the practical deployment of Transformers. Given the emergence of the new-generation architecture of Mamba, this paper proposes a Vision Mamba (VMamba)-based framework for crack segmentation on concrete, asphalt, and masonry surfaces, with high accuracy, generalisation, and less computational complexity. Having 15.6% - 74.5% fewer parameters, the encoder-decoder network integrated with VMamba could obtain up to 2.8% higher mDS than representative CNN-based models while showing about the same performance as Transformer-based models. Moreover, the VMamba-based encoder-decoder network could process high-resolution image input with up to 90.6% lower floating-point operations.

cs.CV↗

arXiv2024

Robust feature knowledge distillation for enhanced performance of lightweight crack segmentation models

Vision-based crack detection faces deployment challenges due to the size of robust models and edge device limitations. These can be addressed with lightweight models trained with knowledge distillation (KD). However, state-of-the-art (SOTA) KD methods compromise anti-noise robustness. This paper develops Robust Feature Knowledge Distillation (RFKD), a framework to improve robustness while retaining the precision of light models for crack segmentation. RFKD distils knowledge from a teacher model's logit layers and intermediate feature maps while leveraging mixed clean and noisy images to transfer robust patterns to the student model, improving its precision, generalisation, and anti-noise performance. To validate the proposed RFKD, a lightweight crack segmentation model, PoolingCrack Tiny (PCT), with only 0.5 M parameters, is also designed and used as the student to run the framework. The results show a significant enhancement in noisy images, with RFKD reaching a 62% enhanced mean Dice score (mDS) compared to SOTA KD methods.

cs.CV↗

arXiv2023

Electric conductivity in non-Hermitian holography

We study the phase structure and charge transport at finite temperature and chemical potential in the non-Hermitian PT-symmetric holographic model of arXiv:1912.06647. The non-Hermitian PT-symmetric deformation is realized by promoting the parameter of a global U(1) symmetry to a complex number. Depending on the strength of the deformation, we find three phases: stable PT-symmetric phase, unstable PT-symmetric phase, and an unstable PT-symmetry broken phase. In the three phases, the square of the condensate and also the spectral weight of the AC conductivity at zero frequency are, respectively, positive, negative, and complex. We check that the Ferrell-Glover-Tinkham sum rule for the AC conductivity holds in all the three phases. We also investigate a complexified U(1) rotor model with PT-symmetric deformation, derive its phase structure and condensation pattern, and find a zero frequency spectral weight analogous to the holographic model.

hep-th↗

arXiv2021

Static charged dilaton black hole cannot be overcharged by gedanken experiments

We consider the new version of the gedanken experiments proposed recently by Sorce and Wald to overcharge a static charged dilaton black hole. First of all, we derive the first-order and second-order perturbation inequalities in the Einstein-Maxwell-dilaton gravitational theory based on the Iyer-Wald formalism. As a result, we find that the weak cosmic censorship conjecture associated with this black hole can be protected after taking into account the second-order perturbation inequality, although violated by the scene without considering this inequality. Therefore, there is no violation of the weak cosmic censorship conjecture around the charged static dilaton black holes in Einstein-Maxwell-dilaton gravity

hep-th↗

arXiv2020

Switchback effect of holographic complexity in multiple-horizon black holes

In this paper, we use the "complexity equals action" (CA) conjecture to explore the switchback effect in the strongly-coupled quantum field theories with finite $N$ and finite coupling effects. In the perspective of holography, this is equivalent to evaluating the CA complexity in a Vaidya geometry equipped with a light shockwave for a higher curvature gravitational theory. Based on the Noether charge formalism of Iyer and Wald, we obtain the slope of the complexity of formation in the small and large time approximations. By circuit analogy, we show that our results concur with the switchback effect of the quantum system. These results show that the switchback effect is a general feature of the CA complexity in stationary black holes and its existence is independent of the explicit gravitational theory as well as spacetime background. From the viewpoint of AdS/CFT, this also implies that the switchback effect is a general feature of the thermofield double state in the strongly-coupled quantum field systems with finite $N$ and finite coupling effects. Moreover, we also illustrate that unlike the late-time complexity growth rate, the counterterm plays an important role in the study of the switchback effect.

hep-th↗

arXiv2019

General Smarr relation and first law of Nutty dyonic black hole

In this paper, we investigate the thermodynamics of the charged NUT black hole with the Misner strings present in the Einstein-Maxwell gravity. We show that the Misner charge $N$ can be obtained by performing Komar integration over the Misner strings in the presence of electric and magnetic charges and the corresponding Smarr relation and first law can also be naturally established based on this geometric perspective. Besides, we demonstrate that the electric and magnetic charges appearing in the first law and Smarr relation are contingent on the special choice of gauge freedom for the electric and magnetic potentials. We re-derive two different versions of thermodynamics according to distinct choices of gauge and furthermore, we try to formulate more general thermodynamic laws under other arbitrary choices of gauge and deduce the conditions necessary to satisfy the Smarr relations and first laws simultaneously.

hep-th↗