arXiv subjects

Xiaofei Liu

Publications and source records attributed to Xiaofei Liu.

18 recordsLinked to original sources

arXiv2026

Performance Evaluation of Delay Tolerant Network Protocols to Improve Nepal Earthquake Rescue Communications

In the fields of disaster rescue and communication in extreme environments, Delay Tolerant Network (DTN) has become an important technology due to its "store-carry-forward" mechanism. Selecting the appropriate routing strategy is of crucial significance for improving the success rate of distress message transmission and reducing delays in material dispatch. We design a pseudo realistic use case of Nepal Kathmandu earthquake rescue based on dynamically changing population distribution model and characteristics of rescue activities in the initial rescue efforts in Nepal Kathmandu earthquakes to conducted the multi criteria two benchmark routing protocols performance analysis in the face of different buffer sizes of the rescue team nodes. We identify multiple real world node groups, including affected residents, rescue teams, drones and ground vehicles and communication models are established according to the movement behaviors of these groups. We analyze the communication of distress messages between edge nodes to obtain performance metrics such as delivered probability, average delay, hop count, and buffer time. By analyzing the multi layer complex data and protocols differences, the research results show the effectiveness of distributed DTN communication methods in the Nepal earthquake rescue use case, reveal existence of trade-offs between transmission reliability and resource utilization of different routing protocols in disaster communication environment and provide a basis for the design of next-generation emergency communication services based on edge nodes.

cs.NI↗

arXiv2026

Detecting Misbehaviors of Large Vision-Language Models by Evidential Uncertainty Quantification

%Large vision-language models (LVLMs) have shown substantial advances in multimodal understanding and generation. However, when presented with incompetent or adversarial inputs, they frequently produce unreliable or even harmful content, such as fact hallucinations or dangerous instructions. This misalignment with human expectations, referred to as \emph{misbehaviors} of LVLMs, raises serious concerns for deployment in critical applications. These misbehaviors are found to stem from epistemic uncertainty, specifically either conflicting internal knowledge or the absence of supporting information. However, existing uncertainty quantification methods, which typically capture only overall epistemic uncertainty, have shown limited effectiveness in identifying such issues. To address this gap, we propose Evidential Uncertainty Quantification (EUQ), a fine-grained method that captures both information conflict and ignorance for effective detection of LVLM misbehaviors. In particular, we interpret features from the model output head as either supporting (positive) or opposing (negative) evidence. Leveraging Evidence Theory, we model and aggregate this evidence to quantify internal conflict and knowledge gaps within a single forward pass. %We extensively evaluate our method across four categories of misbehavior, including hallucinations, jailbreaks, adversarial vulnerabilities, and out-of-distribution (OOD) failures, using state-of-the-art LVLMs, and find that EUQ consistently outperforms strong baselines, showing that hallucinations correspond to high internal conflict and OOD failures to high ignorance. Furthermore, layer-wise evidential uncertainty dynamics analysis helps interpret the evolution of internal representations from a new perspective. The source code is available at https://github.com/HT86159/EUQ.

cs.LG↗

arXiv2025

On-Device Control of Electronic Friction

Friction causes mechanical energy dissipation and material degradation in machinery and devices. While phononic friction is well understood via anharmonic lattice dynamics, the physics of electronic friction remains unclear due to challenges in separating electronic degrees of freedom from phononic ones in experiments and analyzing the non-equilibrium interactions between ionic movement and electronic dynamics in theory. To tackle this problem, we construct a sliding device featuring 2D crystalline interfaces that possess ultra-smooth and minimally interacting surfaces, achieving the state of structural superlubricity with no wear and minimal friction. Using electrical and mechanical controls, we tuned the nature of interfacial electronic coupling and charge densities in materials in an on-device setting, which allows us to disentangle the electron and phonon contributions to friction. Our experimental data and theoretical analysis supported by first-principles calculations demonstrate that electronic friction can well surpass phononic contributions and dominate energy dissipation at structural superlubricity contacts. These findings offer fresh insights into the mechanism of electronic friction and promising opportunities for friction control in device applications.

cond-mat.mes-hall↗

arXiv2024

van der Waals Torque in 2D Materials Induced by Interaction between Many-Body Charge Density Fluctuations

Van der Waals torque determines the relative rotational motion between anisotropic objects, being of relevance to low-dimensional systems. Here we demonstrate a substantial torque between anisotropic two-dimensional materials that arises from the interaction between many-body charge density fluctuations, exceeding by twenty-fold the torque computed with atom-pairwise models. The dependence of torque on the disorientation angle, the positive correlation between torque and in-planar dielectric anisotropy, the linear relation between torque and area, and the decaying torque with increasing separation are rediscovered using the fully atomistic many-body dispersion model. Unlike continuum Casimir-Lifshitz theory, the advantage of the molecular theory relies on describing the collective torque and the effects of atomic details on an equal footing. These findings open an avenue for incorporating quantum fluctuation-induced torque into molecular modeling, being instrumental to the design of nanoelectromechanical systems and the understanding of rotational dynamics of anisotropic layered materials.

physics.comp-ph↗

arXiv2024

Masked adversarial neural network for cell type deconvolution in spatial transcriptomics

Accurately determining cell type composition in disease-relevant tissues is crucial for identifying disease targets. Most existing spatial transcriptomics (ST) technologies cannot achieve single-cell resolution, making it challenging to accurately determine cell types. To address this issue, various deconvolution methods have been developed. Most of these methods use single-cell RNA sequencing (scRNA-seq) data from the same tissue as a reference to infer cell types in ST data spots. However, they often overlook the differences between scRNA-seq and ST data. To overcome this limitation, we propose a Masked Adversarial Neural Network (MACD). MACD employs adversarial learning to align real ST data with simulated ST data generated from scRNA-seq data. By mapping them into a unified latent space, it can minimize the differences between the two types of data. Additionally, MACD uses masking techniques to effectively learn the features of real ST data and mitigate noise. We evaluated MACD on 32 simulated datasets and 2 real datasets, demonstrating its accuracy in performing cell type deconvolution. All code and public datasets used in this paper are available at https://github.com/wenwenmin/MACD and https://zenodo.org/records/12804822.

cs.LG↗

arXiv2023

AlerTiger: Deep Learning for AI Model Health Monitoring at LinkedIn

Data-driven companies use AI models extensively to develop products and intelligent business solutions, making the health of these models crucial for business success. Model monitoring and alerting in industries pose unique challenges, including a lack of clear model health metrics definition, label sparsity, and fast model iterations that result in short-lived models and features. As a product, there are also requirements for scalability, generalizability, and explainability. To tackle these challenges, we propose AlerTiger, a deep-learning-based MLOps model monitoring system that helps AI teams across the company monitor their AI models' health by detecting anomalies in models' input features and output score over time. The system consists of four major steps: model statistics generation, deep-learning-based anomaly detection, anomaly post-processing, and user alerting. Our solution generates three categories of statistics to indicate AI model health, offers a two-stage deep anomaly detection solution to address label sparsity and attain the generalizability of monitoring new models, and provides holistic reports for actionable alerts. This approach has been deployed to most of LinkedIn's production AI models for over a year and has identified several model issues that later led to significant business metric gains after fixing.

cs.LG↗

arXiv2022

Brownian motion of nonlinear oscillator in van der Waals trap

Van der Waals trap, a quantum fluctuation-induced potential characterized by short-range repulsive and long-range attractive forces, is intrinsically nonlinear. This work unveils the nonlinear effects on Brownian oscillators in the van der Waals trap using Langevin dynamics simulations and quasiharmonic approximations. While neither size- nor temperature-dependences of effective natural frequency is important for suspended plates of large areas, smaller ones with broader probability distributions are significantly softened and even a temperature-induced softening is observed. Despite the nonlinearity, the stiffness and the coefficient of friction are tunable by changing the thickness of coating and by modifying the size and the perforation condition of suspended plates, respectively, endowing the quantum trap with flexibilities of building up microscopic mechanical systems and probing near-boundary hydrodynamics.

physics.chem-ph↗

arXiv2021

Security analysis method for practical quantum key distribution with arbitrary encoding schemes

Quantum key distribution (QKD) gradually has become a crucial element of practical secure communication. In different scenarios, the security analysis of genuine QKD systems is complicated. A universal secret key rate calculation method, used for realistic factors such as multiple degrees of freedom encoding, asymmetric protocol structures, equipment flaws, environmental noise, and so on, is still lacking. Based on the correlations of statistical data, we propose a security analysis method without restriction on encoding schemes. This method makes a trade-off between applicability and accuracy, which can effectively analyze various existing QKD systems. We illustrate its ability by analyzing source flaws and a high-dimensional asymmetric protocol. Results imply that our method can give tighter bounds than the Gottesman-Lo-Lütkenhaus-Preskill (GLLP) analysis and is beneficial to analyze protocols with complex encoding structures. Our work has the potential to become a reference standard for the security analysis of practical QKD.

quant-ph↗

arXiv2020

Slope plate of sticky soil granular slope instability based on complex network

The particle discrete element simulation of the instability and failure process of the granular slope accumulator model when the metal plate continues downward is obtained, and the two-dimensional total velocity vector of soil particle velocity and slope slip during the instability and failure of the slope accumulator are obtained. Macro-response processes such as removing the angle of the crack surface and the average velocity in the y-direction of the slope top of the slope accumulation body. Construct a normal force chain undirected network model of the slope accumulation body particles under natural accumulation, and study the location of its slip surface, and The results are compared with the experimental results. Finally, the complex network method is used to analyze the topological characteristics of the contact force chain network of the particles on the slope top of the slope accumulation body, and the average degree, clustering coefficient and average shortest path are obtained during the slope instability of the slope accumulation body. The evolutionary rule of the method is used to verify its accuracy in combination with the strength reduction method. The research results show that the average shortest path can provide a more effective early warning of the instability and failure of slope deposits. A complex network theory is used to study the macro response of the slope deposits and its force chain. The interrelationship between the macroscopic structure of the network provides a new mathematical analysis method for the study of slope instability.

physics.geo-ph↗

arXiv2020

Sustained Coherence Characteristics of Failure Process of Shallow Buried Tunnel under Multiple Explosive Loads based on Persistent Homology

The damage characteristics of a shallow buried tunnel under multiple explosive loads is an important research issue in the design and evaluation of protective engineering. It is of great significance to develop a method for early warning of the safety of the shallow buried features. The discrete element method is used to establish a mechanical model of the shallow buried tunnel. The South Load Equivalent Principle treats blast loads as a series of dynamic forces acting uniformly on the surface. Based on the discrete element method, the dynamic response after each blast load and the damage evolution process of the surrounding rock of the tunnel are obtained. The strength reduction method is used to obtain the surrounding rock of the tunnel. Introduce the theory of continuous homology, and use the mathematical method of continuous homology to quantitatively and qualitatively analyze the failure characteristics of the discrete element model under multiple explosive loads. The results show that the method of continuous homology can accurately reflect the topological characteristics of the surrounding rock of the tunnel The maximum one-dimensional bar code connection radius can effectively warn tunnel instability. This provides a new mathematical method for tunnel safety design and disaster prediction research.

physics.geo-ph↗

arXiv2020

Topological feature study of slope failure process via persistent homology-based machine learning

Using software UDEC to simulate the instability failure process of slope under seismic load, studing the dynamic response of slope failure, obtaining the deformation characteristics and displacement cloud map of slope, then analyzing the instability state of slope by using the theory of persistent homology, generates bar code map and extracts the topological characteristics of slope from bar code map. The topological characteristics corresponding to the critical state of slope instability are found, and the relationship between topological characteristics and instability evolution is established. Finally, it provides a topological research tool for slope failure prediction. The results show that the change of the longest Betti 1 bar code reflects the evolution process of the slope and the law of instability failure. Using discrete element method and persistent homology theory to study the failure characteristics of slope under external load can better understand the failure mechanism of slope, provide theoretical basis for engineering protection, and also provide a new mathematical method for slope safety design and disaster prediction research.

physics.geo-ph↗

arXiv2020

Stress chain analysis of soil particles in slope via persistent homology

The relationship between the macroscopic response of the slope and the macrostructure of the force chain network under the action of the metal plate was studied by the particle discrete element method and the persistent homology. The particle accumulation model was used to simulate the instability process of slope under the continuous downward action of metal plate by the particle discrete element method. The macroscopic responses such as the total velocity vector of the two-dimensional slope deposit, the angle of the slip cracking surface when the slope is unstable, and the average velocity in the y-direction of the slope were studied. Then, the normal force chain undirected network model of the natural accumulation of slope stacking particles was constructed. Finally, the topological characteristics of the particle contact force chain network of the slope top were analyzed by the persistent homology method to obtain the barcode. Finally, the relationship between the instability evolution and the characteristics of persistent homology is established. This research provides a new method for the study of slope instability topology identification. Thus, the instability destruction of slope can be predicted effectively.

cond-mat.soft↗

arXiv2020

Analysis of tunnel failure characteristics under multiple explosion loads based on persistent homology-based machine learning

The study of tunnel failure characteristics under the load of external explosion source is an important problem in tunnel design and protection, in particular, it is of great significance to construct an intelligent topological feature description of the tunnel failure process. The failure characteristics of tunnels under explosive loading are described by using discrete element method and persistent homology-based machine learning. Firstly, the discrete element model of shallow buried tunnel was established in the discrete element software, and the explosive load was equivalent to a series of uniformly distributed loads acting on the surface by Saint-Venant principle, and the dynamic response of the tunnel under multiple explosive loads was obtained through iterative calculation. The topological characteristics of surrounding rock is studied by persistent homology-based machine learning. The geometric, physical and interunit characteristics of the tunnel subjected to explosive loading are extracted, and the nonlinear mapping relationship between the topological quantity of persistent homology, and the failure characteristics of the surrounding rock is established, and the results of the intelligent description of the failure characteristics of the tunnel are obtained. The research shows that the length of the longest Betty 1 bar code is closely related to the stability of the tunnel, which can be used for effective early warning of the tunnel failure, and an intelligent description of the tunnel failure process can be established to provide a new idea for tunnel engineering protection.

cs.CE↗

arXiv2019

Far-field three-dimensional deep-subwavelength focal spot with azimuthal polarization

This work focuses on the generation of far-field super-resolved pure-azimuthal focal field based on the fast Fourier transform. A self-designed differential filter is first pioneered to robustly reconfigure a doughnut-shaped azimuthal focal field into a bright one with a sub-wavelength lateral scale (0.392λ), which offers a 27.3% reduction ratio relative to that of tightly focused azimuthal polarization modulated by a spiral phase plate. By further uniting the versatile differential filter with spatially shifted beam approach, in addition to allowing for an extremely sharper focal spot, whose size is in turn reduced to 0.228λ and 0.286λ in the transverse as well as axial directions, the parasitic sidelobes are also lowered to an inessential level (< 20%), thereby enabling an excellent three-dimensional deep-subwavelength focal field (λ3/128). The relevant phase profiles are further exhibited to unravel the annihilation of field singularity and locally linear (i.e. azimuthal) polarization. Our scheme opens a promising route toward efficiently steer and tailor the redistribution of the focal field.

physics.optics↗

arXiv2017

ATRank: An Attention-Based User Behavior Modeling Framework for Recommendation

A user can be represented as what he/she does along the history. A common way to deal with the user modeling problem is to manually extract all kinds of aggregated features over the heterogeneous behaviors, which may fail to fully represent the data itself due to limited human instinct. Recent works usually use RNN-based methods to give an overall embedding of a behavior sequence, which then could be exploited by the downstream applications. However, this can only preserve very limited information, or aggregated memories of a person. When a downstream application requires to facilitate the modeled user features, it may lose the integrity of the specific highly correlated behavior of the user, and introduce noises derived from unrelated behaviors. This paper proposes an attention based user behavior modeling framework called ATRank, which we mainly use for recommendation tasks. Heterogeneous user behaviors are considered in our model that we project all types of behaviors into multiple latent semantic spaces, where influence can be made among the behaviors via self-attention. Downstream applications then can use the user behavior vectors via vanilla attention. Experiments show that ATRank can achieve better performance and faster training process. We further explore ATRank to use one unified model to predict different types of user behaviors at the same time, showing a comparable performance with the highly optimized individual models.

cs.AI↗

arXiv2017

Negative thermal expansion and metallophilicity in Cu$_3$[Co(CN)$_6$]

We report the synthesis and structural characterisation of the molecular framework copper(I) hexacyanocobaltate(III), Cu$_3$[Co(CN)$_6$], which we find to be isostructural to H$_3$[Co(CN)$_6$] and the colossal negative thermal expansion material Ag$_3$[Co(CN)$_6$]. Using synchrotron X-ray powder diffraction measurements, we find strong positive and negative thermal expansion behaviour respectively perpendicular and parallel to the trigonal crystal axis: $α_a$ = 25.4(5)\,MK$^{-1}$ and $α_c$ = $-$43.5(8)\,MK$^{-1}$. These opposing effects collectively result in a volume expansivity $α_V$ = 7.4(11)\,MK$^{-1}$ that is remarkably small for an anisotropic molecular framework. This thermal response is discussed in the context of the behaviour of the analogous H- and Ag-containing systems. We make use of density-functional theory with many-body dispersion interactions (DFT+MBD) to demonstrate that Cu$\ldots$Cu metallophilic (`cuprophilic') interactions are significantly weaker in Cu$_3$[Co(CN)$_6$] than Ag$\ldots$Ag interactions in Ag$_3$[Co(CN)$_6$], but that this lowering of energy scale counterintuitively translates to a more moderate---rather than enhanced---degree of structural flexibility. The same conclusion is drawn from consideration of a simple lattice dynamical model, which we also present here. Our results demonstrate that strong interactions can actually be exploited in the design of ultra-responsive materials if those interactions are set up to act in tension.

cond-mat.mtrl-sci↗

arXiv2015

A Graph Theoretic Characterization of Perfect Attackability and Detection in Distributed Control Systems

This paper is concerned with the analysis and design of secure Distributed Control Systems in the face of integrity attacks on sensors and controllers by external attackers or insiders. In general a DCS consists of many heterogenous components and agents including sensors, actuators, controllers. Due to its distributed nature, some agents may start misbehaving to disrupt the system. This paper first reviews necessary and sufficient conditions for deterministic detection of integrity attacks carried out by any number of malicious agents, based on the concept of left invertibility of structural control systems. It then develops a notion equivalent to structural left invertibility in terms of vertex separators of a graph. This tool is then leveraged to design minimal communication networks for DCSs, which ensure that an adversary cannot generate undetectable attacks. Numerical examples are included to illustrate these results.

eess.SY↗

arXiv2015

Minimum Sensor Placement for Robust Observability of Structured Complex Networks

This paper addresses problems on the robust structural design of complex networks. More precisely, we address the problem of deploying the minimum number of dedicated sensors, i.e., those measuring a single state variable, that ensure the network to be structurally observable under disruptive scenarios. The disruptive scenarios considered are as follows: (i) the malfunction/loss of one arbitrary sensor, and (ii) the failure of connection (either unidirectional or bidirectional communication) between a pair of agents. First, we show these problems to be NP-hard, which implies that efficient algorithms to determine a solution are unlikely to exist. Secondly, we propose an intuitive two step approach: (1) we achieve an arbitrary minimum sensor placement ensuring structural observability; (2) we develop a sequential process to find minimum number of additional sensors required for robust observability. This step can be solved by recasting it as a weighted set covering problem. Although this is known to be an NP-hard problem, feasible approximations can be determined in polynomial-time that can be used to obtain feasible approximations to the robust structural design problems with optimality guarantees.

math.OC↗