Selected work

Representative Papers

Out of Sync, Out of Sight: Phantom State Attacks against IIoT Intrusion Detection

Oct 01, 2026

This study addresses the vulnerability of Industrial Internet of Things (IIoT) intrusion detection systems that rely on time-window aggregation to temporal synchronization attacks. We propose the Phantom State Attack (PSA), which injects bounded temporal drifts to push observations toward window boundaries, thereby inducing erroneous state reconstruction. Operating in a zero-query passive mode, PSA requires neither packet tampering nor model fitting, exploiting solely the vulnerabilities inherent in temporal aggregation. Analytical evaluations under both sliding and tumbling window configurations, along with empirical assessments on the ToN-IoT and CIC datasets, demonstrate that PSA significantly degrades detection rates for targeted traffic. Compared to baseline methods, the proposed attack achieves superior efficiency and stealthiness, revealing a novel attack surface for IIoT security.

0 citationsRead paper

MMM: Clustering Multivariate Longitudinal Mixed-type Data

Sep 15, 2025

Clustering multivariate longitudinal data with mixed variable types—continuous, ordinal, binary, nominal, and count—poses significant challenges due to heterogeneous individual trajectories, complex inter-variable dependencies, and temporal dynamics. Method: We propose the MMM (Mixed-type Multivariate Longitudinal) model, which reformulates the data as a three-way tensor and jointly models individual heterogeneity, cross-variable associations, and temporal dependence within a latent variable space. MMM extends the matrix-normal mixture framework to mixed-type longitudinal settings, explicitly relaxing the conventional conditional independence assumption. Non-continuous variables are handled via latent variable mappings, while parameter inference is performed using a matrix-variate normal mixture formulation coupled with an MCMC-EM algorithm. Contribution/Results: MMM is the first method to unify multidimensional dependency structures in mixed-type longitudinal clustering. Experiments demonstrate substantially higher clustering accuracy on synthetic benchmarks versus state-of-the-art baselines; on real financial longitudinal data, MMM yields both high interpretability and robust analytical performance.

0 citationsRead paper

Recent publications

Latest Papers

Out of Sync, Out of Sight: Phantom State Attacks against IIoT Intrusion Detection

Oct 01, 2026

This study addresses the vulnerability of Industrial Internet of Things (IIoT) intrusion detection systems that rely on time-window aggregation to temporal synchronization attacks. We propose the Phantom State Attack (PSA), which injects bounded temporal drifts to push observations toward window boundaries, thereby inducing erroneous state reconstruction. Operating in a zero-query passive mode, PSA requires neither packet tampering nor model fitting, exploiting solely the vulnerabilities inherent in temporal aggregation. Analytical evaluations under both sliding and tumbling window configurations, along with empirical assessments on the ToN-IoT and CIC datasets, demonstrate that PSA significantly degrades detection rates for targeted traffic. Compared to baseline methods, the proposed attack achieves superior efficiency and stealthiness, revealing a novel attack surface for IIoT security.

0 citationsRead paper

MMM: Clustering Multivariate Longitudinal Mixed-type Data

Sep 15, 2025

Clustering multivariate longitudinal data with mixed variable types—continuous, ordinal, binary, nominal, and count—poses significant challenges due to heterogeneous individual trajectories, complex inter-variable dependencies, and temporal dynamics. Method: We propose the MMM (Mixed-type Multivariate Longitudinal) model, which reformulates the data as a three-way tensor and jointly models individual heterogeneity, cross-variable associations, and temporal dependence within a latent variable space. MMM extends the matrix-normal mixture framework to mixed-type longitudinal settings, explicitly relaxing the conventional conditional independence assumption. Non-continuous variables are handled via latent variable mappings, while parameter inference is performed using a matrix-variate normal mixture formulation coupled with an MCMC-EM algorithm. Contribution/Results: MMM is the first method to unify multidimensional dependency structures in mixed-type longitudinal clustering. Experiments demonstrate substantially higher clustering accuracy on synthetic benchmarks versus state-of-the-art baselines; on real financial longitudinal data, MMM yields both high interpretability and robust analytical performance.

0 citationsRead paper