arXiv subjects

Kai Zhou

Publications and source records attributed to Kai Zhou.

At least 73 records · Page 4Linked to original sources

arXiv2025

Robust Graph Contrastive Learning with Information Restoration

The graph contrastive learning (GCL) framework has gained remarkable achievements in graph representation learning. However, similar to graph neural networks (GNNs), GCL models are susceptible to graph structural attacks. As an unsupervised method, GCL faces greater challenges in defending against adversarial attacks. Furthermore, there has been limited research on enhancing the robustness of GCL. To thoroughly explore the failure of GCL on the poisoned graphs, we investigate the detrimental effects of graph structural attacks against the GCL framework. We discover that, in addition to the conventional observation that graph structural attacks tend to connect dissimilar node pairs, these attacks also diminish the mutual information between the graph and its representations from an information-theoretical perspective, which is the cornerstone of the high-quality node embeddings for GCL. Motivated by this theoretical insight, we propose a robust graph contrastive learning framework with a learnable sanitation view that endeavors to sanitize the augmented graphs by restoring the diminished mutual information caused by the structural attacks. Additionally, we design a fully unsupervised tuning strategy to tune the hyperparameters without accessing the label information, which strictly coincides with the defender's knowledge. Extensive experiments demonstrate the effectiveness and efficiency of our proposed method compared to competitive baselines.

cs.LG↗

arXiv2025

Thermodynamics of Heavy Quarkonium in a Bayesian Holographic QCD model

Leveraging high-precision lattice QCD data on the equation of state and baryon number susceptibility at vanishing chemical potential, we construct a Bayesian holographic QCD model and systematically analyze the thermodynamic properties of heavy quarkonium in QCD matter under varying temperatures and chemical potentials. We compute the quark-antiquark interquark distance, potential energy, entropy, binding energy, and internal energy. We present detailed posterior distribution results of the thermodynamic quantities of heavy quarkonium, including maximum a posteriori (MAP) value estimates and 95\% confidence levels (CL). Through numerical simulations and theoretical analysis, we find that increasing temperature and chemical potential decrease the quark distance, thereby facilitating the dissociation of heavy quarkonium and leading to suppressed potential energy. The increase in temperature and chemical potential also raise the entropy and entropy force, further accelerating the dissociation of heavy quarkonium. The calculated results of binding energy indicate that higher temperature and chemical potential enhance the tendency of heavy quarkonium to dissociate into free quarks. Internal energy also increases with rising temperature and chemical potential. These findings provide significant theoretical insights into the properties of strongly interacting matter under extreme conditions and lay a solid foundation for the interpretation and validation of future experimental data. Finally, we also present the results for the free energy, entropy, and internal energy of single quark.

hep-ph↗

arXiv2025

Hierarchical knowledge guided fault intensity diagnosis of complex industrial systems

Fault intensity diagnosis (FID) plays a pivotal role in monitoring and maintaining mechanical devices within complex industrial systems. As current FID methods are based on chain of thought without considering dependencies among target classes. To capture and explore dependencies, we propose a hierarchical knowledge guided fault intensity diagnosis framework (HKG) inspired by the tree of thought, which is amenable to any representation learning methods. The HKG uses graph convolutional networks to map the hierarchical topological graph of class representations into a set of interdependent global hierarchical classifiers, where each node is denoted by word embeddings of a class. These global hierarchical classifiers are applied to learned deep features extracted by representation learning, allowing the entire model to be end-to-end learnable. In addition, we develop a re-weighted hierarchical knowledge correlation matrix (Re-HKCM) scheme by embedding inter-class hierarchical knowledge into a data-driven statistical correlation matrix (SCM) which effectively guides the information sharing of nodes in graphical convolutional neural networks and avoids over-smoothing issues. The Re-HKCM is derived from the SCM through a series of mathematical transformations. Extensive experiments are performed on four real-world datasets from different industrial domains (three cavitation datasets from SAMSON AG and one existing publicly) for FID, all showing superior results and outperform recent state-of-the-art FID methods.

cs.AI↗

arXiv2025

Revisiting Privacy-Utility Trade-off for DP Training with Pre-existing Knowledge

Differential privacy (DP) provides a provable framework for protecting individuals by customizing a random mechanism over a privacy-sensitive dataset. Deep learning models have demonstrated privacy risks in model exposure as an established learning model unintentionally records membership-level privacy leakage. Differentially private stochastic gradient descent (DP-SGD) has been proposed to safeguard training individuals by adding random Gaussian noise to gradient updates in the backpropagation. Researchers identify that DP-SGD causes utility loss since the injected homogeneous noise can alter the gradient updates calculated at each iteration. Namely, all elements in the gradient are contaminated regardless of their importance in updating model parameters. In this work, we argue that the utility can be optimized by involving the heterogeneity of the the injected noise. Consequently, we propose a generic differential privacy framework with heterogeneous noise (DP-Hero) by defining a heterogeneous random mechanism to abstract its property. The insight of DP-Hero is to leverage the knowledge encoded in the previously trained model to guide the subsequent allocation of noise heterogeneity, thereby leveraging the statistical perturbation and achieving enhanced utility. Atop DP-Hero, we instantiate a heterogeneous version of DP-SGD, and further extend it to federated training. We conduct comprehensive experiments to verify and explain the effectiveness of the proposed DP-Hero, showing improved training accuracy compared with state-of-the-art works. Broadly, we shed light on improving the privacy-utility space by learning the noise guidance from the pre-existing leaked knowledge encoded in the previously trained model, showing a different perspective of understanding the utility-improved DP training.

cs.CR↗

arXiv2025

Multi-Grained Temporal-Spatial Graph Learning for Stable Traffic Flow Forecasting

Time-evolving traffic flow forecasting are playing a vital role in intelligent transportation systems and smart cities. However, the dynamic traffic flow forecasting is a highly nonlinear problem with complex temporal-spatial dependencies. Although the existing methods has provided great contributions to mine the temporal-spatial patterns in the complex traffic networks, they fail to encode the globally temporal-spatial patterns and are prone to overfit on the pre-defined geographical correlations, and thus hinder the model's robustness on the complex traffic environment. To tackle this issue, in this work, we proposed a multi-grained temporal-spatial graph learning framework to adaptively augment the globally temporal-spatial patterns obtained from a crafted graph transformer encoder with the local patterns from the graph convolution by a crafted gated fusion unit with residual connection techniques. Under these circumstances, our proposed model can mine the hidden global temporal-spatial relations between each monitor stations and balance the relative importance of local and global temporal-spatial patterns. Experiment results demonstrate the strong representation capability of our proposed method and our model consistently outperforms other strong baselines on various real-world traffic networks.

cs.LG↗

arXiv2025

Zero-Shot Skeleton-Based Action Recognition With Prototype-Guided Feature Alignment

Zero-shot skeleton-based action recognition aims to classify unseen skeleton-based human actions without prior exposure to such categories during training. This task is extremely challenging due to the difficulty in generalizing from known to unknown actions. Previous studies typically use two-stage training: pre-training skeleton encoders on seen action categories using cross-entropy loss and then aligning pre-extracted skeleton and text features, enabling knowledge transfer to unseen classes through skeleton-text alignment and language models' generalization. However, their efficacy is hindered by 1) insufficient discrimination for skeleton features, as the fixed skeleton encoder fails to capture necessary alignment information for effective skeleton-text alignment; 2) the neglect of alignment bias between skeleton and unseen text features during testing. To this end, we propose a prototype-guided feature alignment paradigm for zero-shot skeleton-based action recognition, termed PGFA. Specifically, we develop an end-to-end cross-modal contrastive training framework to improve skeleton-text alignment, ensuring sufficient discrimination for skeleton features. Additionally, we introduce a prototype-guided text feature alignment strategy to mitigate the adverse impact of the distribution discrepancy during testing. We provide a theoretical analysis to support our prototype-guided text feature alignment strategy and empirically evaluate our overall PGFA on three well-known datasets. Compared with the top competitor SMIE method, our PGFA achieves absolute accuracy improvements of 22.96%, 12.53%, and 18.54% on the NTU-60, NTU-120, and PKU-MMD datasets, respectively.

cs.CV↗

arXiv2025

Defense-as-a-Service: Black-box Shielding against Backdoored Graph Models

With the trend of large graph learning models, business owners tend to employ a model provided by a third party to deliver business services to users. However, these models might be backdoored, and malicious users can submit trigger-embedded inputs to manipulate the model predictions. Current graph backdoor defenses have several limitations: 1) depending on model-related details, 2) requiring additional model fine-tuning, and 3) relying upon extra explainability tools, all of which are infeasible under stringent privacy policies. To address those limitations, we propose GraphProt, which allows resource-constrained business owners to rely on third parties to avoid backdoor attacks on GNN-based graph classifiers. Our GraphProt is model-agnostic and only relies on the input graph. The key insight is to leverage subgraph information for prediction, thereby mitigating backdoor effects induced by triggers. GraphProt comprises two components: clustering-based trigger elimination and robust subgraph ensemble. Specifically, we first propose feature-topology clustering that aims to remove most of the anomalous subgraphs (triggers). Moreover, we design subgraph sampling strategies based on feature-topology clustering to build a robust classifier via majority vote. Experimental results across three backdoor attacks and six benchmark datasets demonstrate that GraphProt significantly reduces the backdoor attack success rate while preserving the model accuracy on regular graph classification tasks.

cs.LG↗

arXiv2025

Exploring Transport Properties of Quark-Gluon Plasma in Flavor-Dependent Systems with a Holographic Model

Based on the holographic model, which incorporates the equation of state (EoS) and baryon number susceptibility for different flavors, we calculate the drag force, jet quenching parameter, and diffusion coefficient of the heavy quark at finite temperature and chemical potential. The holographic results for the diffusion coefficient align with lattice data for $N_f = 0$ and $N_f = 2+1$, falling within their error margins. The holographic diffusion coefficient for heavy quark in the systems of different flavors \textcolor{red}{is compatible with estimates from ALICE data.} The jet quenching parameter in our model demonstrates strong consistency \textcolor{red}{with the estimations obtained from Bayesian analysis of data from both RHIC and LHC for different flavors.} We can confirm the model provides a good description of the transport properties of QGP. The work reinforces the potential of bottom-up holographic model in advancing our understanding of transport properties of hot and dense quark-gluon plasma.

hep-ph↗

arXiv2025

Collapse of a quantum vortex in an attractive two-dimensional Bose gas

We experimentally and numerically study the collapse dynamics of a quantum vortex in a two-dimensional atomic superfluid following a fast interaction ramp from repulsion to attraction. We find the conditions and time scales for a superfluid vortex to radially converge into a quasi-stationary density profile, demonstrating the spontaneous formation of a vortex soliton-like structure in an atomic Bose gas. We record an emergent self-similar dynamics caused by an azimuthal modulational instability, which amplifies initial density perturbations and leads to the eventual splitting of a solitonic ring profile or direct fragmentation of a superfluid into disordered, but roughly circular arrays of Townes soliton-like wavepackets. These dynamics are qualitatively reproduced by simulations based on the Gross-Pitaevskii equation. However, a discrepancy in the magnitude of amplified density fluctuations predicted by our mean-field analysis suggests the presence of effects beyond the mean-field approximation. Our study sets the stage for exploring out-of-equilibrium dynamics of vortex quantum matter quenched to attractive interactions and their universal characteristics.

cond-mat.quant-gas↗

arXiv2025

LT-PINN: Lagrangian Topology-conscious Physics-informed Neural Network for Boundary-focused Engineering Optimization

Physics-informed neural networks (PINNs) have emerged as a powerful meshless tool for topology optimization, capable of simultaneously determining optimal topologies and physical solutions. However, conventional PINNs rely on density-based topology descriptions, which necessitate manual interpolation and limit their applicability to complex geometries. To address this, we propose Lagrangian topology-conscious PINNs (LT-PINNs), a novel framework for boundary-focused engineering optimization. By parameterizing the control variables of topology boundary curves as learnable parameters, LT-PINNs eliminate the need for manual interpolation and enable precise boundary determination. We further introduce specialized boundary condition loss function and topology loss function to ensure sharp and accurate boundary representations, even for intricate topologies. The accuracy and robustness of LT-PINNs are validated via two types of partial differential equations (PDEs), including elastic equation with Dirichlet boundary conditions and Laplace's equation with Neumann boundary conditions. Furthermore, we demonstrate effectiveness of LT-PINNs on more complex time-dependent and time-independent flow problems without relying on measurement data, and showcase their engineering application potential in flow velocity rearrangement, transforming a uniform upstream velocity into a sine-shaped downstream profile. The results demonstrate (1) LT-PINNs achieve substantial reductions in relative L2 errors compared with the state-of-art density topology-oriented PINNs (DT-PINNs), (2) LT-PINNs can handle arbitrary boundary conditions, making them suitable for a wide range of PDEs, and (3) LT-PINNs can infer clear topology boundaries without manual interpolation, especially for complex topologies.

cs.LG↗

arXiv2025

Optimal estimation of three parallel spins with genuine and restricted collective measurements

Collective measurements on identical and independent quantum systems can offer advantages in information extraction compared with individual measurements. However, little is known about the distinction between restricted collective measurements and genuine collective measurements in the multipartite setting. In this work we establish a rigorous performance gap based on a simple and old estimation problem, the estimation of a random spin state given three parallel spins. Notably, we derive an analytical formula for the maximum estimation fidelity of biseparable measurements and clarify its fidelity gap from genuine collective measurements. Moreover, we clarify the structure of optimal biseparable measurements. It turns out that the maximum estimation fidelity can be achieved by two- and one-copy measurements assisted by one-way communication in one direction, but not the other way. Our work reveals a rich landscape of multipartite nonclassicality in quantum measurements instead of quantum states and is expected to trigger further studies.

quant-ph↗

arXiv2025

Experimental Realization of Genuine Three-copy Collective Measurements for Optimal Information Extraction

Nonclassical phenomena tied to entangled states are the focus of foundational studies and powerful resources in many applications. By contrast, the counterparts in quantum measurements are still poorly understood. Notably, genuine multipartite nonclassicality is barely discussed, let alone its experimental realization. Here we experimentally demonstrate the power of genuine tripartite nonclassicality in quantum measurements based on a simple estimation problem. To this end we realize an optimal genuine three-copy collective measurement via a nine-step two-dimensional photonic quantum walk with 30 elaborately designed coin operators. Then we realize an optimal estimation protocol and achieve an unprecedented high estimation fidelity, which can beat all strategies based on restricted collective measurements by more than 11 standard deviations. These results clearly demonstrate that genuine collective measurements can extract more information than local measurements and restricted collective measurements. Our work opens the door for exploring genuine multipartite nonclassical measurements and their power in quantum information processing.

quant-ph↗

arXiv2025

Real-Time Cascade Mitigation in Power Systems Using Influence Graph Improved by Reinforcement Learning

Despite high reliability, modern power systems with growing renewable penetration face an increasing risk of cascading outages. Real-time cascade mitigation requires fast, complex operational decisions under uncertainty. In this work, we extend the influence graph into a Markov decision process model (MDP) for real-time mitigation of cascading outages in power transmission systems, accounting for uncertainties in generation, load, and initial contingencies. The MDP includes a do-nothing action to allow for conservative decision-making and is solved using reinforcement learning. We present a policy gradient learning algorithm initialized with a policy corresponding to the unmitigated case and designed to handle invalid actions. The proposed learning method converges faster than the conventional algorithm. Through careful reward design, we learn a policy that takes conservative actions without deteriorating system conditions. The model is validated on the IEEE 14-bus and IEEE 118-bus systems. The results show that proactive line disconnections can effectively reduce cascading risk, and certain lines consistently emerge as critical in mitigating cascade propagation.

physics.soc-ph↗

arXiv2025

Towards constraining QCD phase transitions in neutron star interiors: Bayesian Inference with TOV linear response analysis

The potential hadron-to-quark phase transition in neutron stars has not been fully understood as the property of cold, dense, and strongly interacting matter cannot be theoretically described by the first-principle perturbative calculations, nor have they been systematically measured through terrestrial low-to-intermediate energy heavy-ion experiments. Given the Tolman--Oppenheimer--Volkoff (TOV) equations, the equation of state (EoS) of the neutron star (NS) matter can be constrained by the observations of NS mass, radius, and tidal deformability. However, large observational uncertainties and the limited number of observations currently make it challenging to strictly reconstruct the EoS, especially to identify interesting features such as a strong first-order phase transition. In this work, we study the dependency of reconstruction quality of the phase transition on the number of NS observations of mass and radius as well as their uncertainty, based on a fiducial EoS. We conquer this challenging problem by constructing a neural network, which allows one to parametrize the EoS with minimum model-dependency, and by devising an algorithm of parameter optimization based on the analytical linear response analysis of the TOV equations. This work may pave the way for the understanding of the phase transition features in NSs using future $x$-ray and gravitational wave measurements.

nucl-th↗

arXiv2025

Inferring the Equation of State from Neutron Star Observables via Machine Learning

We have conducted an extensive study using a diverse set of equations of state (EoSs) to uncover strong relationships between neutron star (NS) observables and the underlying EoS parameters using symbolic regression method. These EoS models, derived from a mix of agnostic and physics-based approaches, considered neutron stars composed of nucleons, hyperons, and other exotic degrees of freedom in beta equilibrium. The maximum mass of a NS is found to be strongly correlated with the pressure and baryon density at an energy density of approximately 800 MeV.fm$^{-3}$. We have also demonstrated that the EoS can be expressed as a function of radius and tidal deformability within the NS mass range 1-2$M_\odot$. These insights offer a promising and efficient framework to decode the dense matter EoS directly from the accurate knowledge of NS observables.

nucl-th↗

arXiv2025

Unifying Physics- and Data-Driven Modeling via Novel Causal Spatiotemporal Graph Neural Network for Interpretable Epidemic Forecasting

Accurate epidemic forecasting is crucial for effective disease control and prevention. Traditional compartmental models often struggle to estimate temporally and spatially varying epidemiological parameters, while deep learning models typically overlook disease transmission dynamics and lack interpretability in the epidemiological context. To address these limitations, we propose a novel Causal Spatiotemporal Graph Neural Network (CSTGNN), a hybrid framework that integrates a Spatio-Contact SIR model with Graph Neural Networks (GNNs) to capture the spatiotemporal propagation of epidemics. Inter-regional human mobility exhibits continuous and smooth spatiotemporal patterns, leading to adjacent graph structures that share underlying mobility dynamics. To model these dynamics, we employ an adaptive static connectivity graph to represent the stable components of human mobility and utilize a temporal dynamics model to capture fluctuations within these patterns. By integrating the adaptive static connectivity graph with the temporal dynamics graph, we construct a dynamic graph that encapsulates the comprehensive properties of human mobility networks. Additionally, to capture temporal trends and variations in infectious disease spread, we introduce a temporal decomposition model to handle temporal dependence. This model is then integrated with a dynamic graph convolutional network for epidemic forecasting. We validate our model using real-world datasets at the provincial level in China and the state level in Germany. Extensive studies demonstrate that our method effectively models the spatiotemporal dynamics of infectious diseases, providing a valuable tool for forecasting and intervention strategies. Furthermore, analysis of the learned parameters offers insights into disease transmission mechanisms, enhancing the interpretability and practical applicability of our model.

cs.LG↗

arXiv2025

Efficient Calibration for RRAM-based In-Memory Computing using DoRA

Resistive In-Memory Computing (RIMC) offers ultra-efficient computation for edge AI but faces accuracy degradation due to RRAM conductance drift over time. Traditional retraining methods are limited by RRAM's high energy consumption, write latency, and endurance constraints. We propose a DoRA-based calibration framework that restores accuracy by compensating influential weights with minimal calibration parameters stored in SRAM, leaving RRAM weights untouched. This eliminates in-field RRAM writes, ensuring energy-efficient, fast, and reliable calibration. Experiments on RIMC-based ResNet50 (ImageNet-1K) demonstrate 69.53% accuracy restoration using just 10 calibration samples while updating only 2.34% of parameters.

cs.AR↗

arXiv2025

On learning higher-order cumulants in diffusion models

To analyse how diffusion models learn correlations beyond Gaussian ones, we study the behaviour of higher-order cumulants, or connected n-point functions, under both the forward and backward process. We derive explicit expressions for the moment- and cumulant-generating functionals, in terms of the distribution of the initial data and properties of forward process. It is shown analytically that during the forward process higher-order cumulants are conserved in models without a drift, such as the variance-expanding scheme, and that therefore the endpoint of the forward process maintains nontrivial correlations. We demonstrate that since these correlations are encoded in the score function, higher-order cumulants are learnt in the backward process, also when starting from a normal prior. We confirm our analytical results in an exactly solvable toy model with nonzero cumulants and in scalar lattice field theory.

hep-lat↗