Selected work

Representative Papers

Hybrid Deep Learning for Traceability and Classification of Industrial Slate Tiles

Jul 06, 2026

This study addresses the high cost and error-proneness of manual inspection of natural slate tiles, which exhibit significant visual variability. To tackle this challenge, the authors propose a lightweight hybrid deep learning framework that, for the first time, integrates XFeat and MobileNetV3 within a unified architecture. The model employs a dual-branch feature sharing and fusion mechanism to jointly optimize instance re-identification and quarry origin classification. Specifically, XFeat coupled with LightGlue enables high-precision image matching, while MobileNetV3 handles source classification. Evaluated on a newly curated industrial dataset comprising 2,610 tile images, the proposed method achieves a 15.4% improvement in instance matching AUC and a 10.9% increase in classification accuracy over the standard MobileNetV3 baseline.

0 citationsRead paper

Understanding How Humans Inject Knowledge into Machine Learning Workflows through Visual Analytics

Jul 01, 2026

This study investigates the mechanisms and pathways through which human domain knowledge is integrated into machine learning (ML) workflows via visual analytics. Building upon a systematic review of over 200 VIS4ML papers, the authors develop a coding framework encompassing four dimensions: machine learning, visualization, interaction, and action. By synthesizing perspectives from model construction and information-theoretic cost–benefit analysis, they propose the first unified explanatory framework for knowledge injection in ML. The work elucidates the pivotal role of interactive visualization in optimizing ML workflows and systematically maps the multidimensional pathways through which human expertise is incorporated. This contribution provides both theoretical grounding and empirical foundations for advancing research and practice in the VIS4ML community.

0 citationsRead paper

From Idea to Prototype in an Afternoon: Scaffolded, AI-Assisted Rapid VA Prototyping

Jun 30, 2026

This work addresses the protracted development cycles in traditional visual analytics (VA) prototyping that hinder rapid validation of novel ideas. The authors propose a scaffolded, AI-assisted development paradigm centered on the Artifact–Transform Workflow Language (ATWL) as a structured framework, integrating large language model–driven AI assistants with targeted expert interventions to efficiently construct high-quality VA prototypes within hours. The approach successfully instantiated innovative visual designs such as “soft Pareto fronts” and “constellation” groupings. Controlled experiments further revealed the critical influence of scaffolding design, timing of human-AI collaboration, and methods of knowledge injection on prototype quality, leading the authors to advocate for a taxonomy of knowledge expression in human-AI collaborative systems.

0 citationsRead paper

A Hybrid Intrusion Detection System for Electric Vehicle Charging Infrastructure

Jun 22, 2026

This study addresses the heightened security risks faced by electric vehicle (EV) charging infrastructure upon integration into smart grids, particularly the threat of cross-domain cyber-physical attacks that existing intrusion detection systems inadequately cover. To counter this challenge, the authors propose a novel two-tier collaborative hybrid intrusion detection system that synergistically integrates network-based (NIDS) and host-based (HIDS) approaches. This framework achieves, for the first time in EV charging scenarios, deep fusion of dual-source data to effectively detect sophisticated composite attacks spanning both cyber and physical layers. Experimental evaluation on the CICEVSE2024 dataset demonstrates that the NIDS component attains a 99.99% accuracy in identifying network-level attacks, while the HIDS component achieves an 83.47% accuracy in detecting false data injection attacks (FDIA), cryptojacking, backdoors, and most denial-of-service (DoS) and reconnaissance attacks—significantly outperforming standalone detection methods.

0 citationsRead paper

ICAN-Deploy: Identity-Stable Canary Deployment for Safety-Critical Embodied Agents

May 27, 2026

This work addresses the challenge of repeated re-certification in traditional canary deployments for safety-critical embodied intelligent agents, which arises from changes to the system’s cryptographic identity. The authors propose ICAN-Deploy, a middleware that decouples immutable capability names—hashed to serve as stable identities—from mutable capability versions, thereby preserving identity hashes throughout the canary window and enabling, for the first time, identity-stable canary deployment. The approach integrates a state machine design, a runtime governance layer, AST-based static analysis, closed-form proofs, and TLA+ model checking to guarantee safety and correctness. Empirical validation on a Franka Panda robotic arm in MuJoCo across 100 deployments demonstrates zero identity drift, with entry latency falling within a 95% BCa confidence interval of [1.52, 2.01] milliseconds.

0 citationsRead paper

Recent publications

Latest Papers

Hybrid Deep Learning for Traceability and Classification of Industrial Slate Tiles

Jul 06, 2026

This study addresses the high cost and error-proneness of manual inspection of natural slate tiles, which exhibit significant visual variability. To tackle this challenge, the authors propose a lightweight hybrid deep learning framework that, for the first time, integrates XFeat and MobileNetV3 within a unified architecture. The model employs a dual-branch feature sharing and fusion mechanism to jointly optimize instance re-identification and quarry origin classification. Specifically, XFeat coupled with LightGlue enables high-precision image matching, while MobileNetV3 handles source classification. Evaluated on a newly curated industrial dataset comprising 2,610 tile images, the proposed method achieves a 15.4% improvement in instance matching AUC and a 10.9% increase in classification accuracy over the standard MobileNetV3 baseline.

0 citationsRead paper

Understanding How Humans Inject Knowledge into Machine Learning Workflows through Visual Analytics

Jul 01, 2026

This study investigates the mechanisms and pathways through which human domain knowledge is integrated into machine learning (ML) workflows via visual analytics. Building upon a systematic review of over 200 VIS4ML papers, the authors develop a coding framework encompassing four dimensions: machine learning, visualization, interaction, and action. By synthesizing perspectives from model construction and information-theoretic cost–benefit analysis, they propose the first unified explanatory framework for knowledge injection in ML. The work elucidates the pivotal role of interactive visualization in optimizing ML workflows and systematically maps the multidimensional pathways through which human expertise is incorporated. This contribution provides both theoretical grounding and empirical foundations for advancing research and practice in the VIS4ML community.

0 citationsRead paper

From Idea to Prototype in an Afternoon: Scaffolded, AI-Assisted Rapid VA Prototyping

Jun 30, 2026

This work addresses the protracted development cycles in traditional visual analytics (VA) prototyping that hinder rapid validation of novel ideas. The authors propose a scaffolded, AI-assisted development paradigm centered on the Artifact–Transform Workflow Language (ATWL) as a structured framework, integrating large language model–driven AI assistants with targeted expert interventions to efficiently construct high-quality VA prototypes within hours. The approach successfully instantiated innovative visual designs such as “soft Pareto fronts” and “constellation” groupings. Controlled experiments further revealed the critical influence of scaffolding design, timing of human-AI collaboration, and methods of knowledge injection on prototype quality, leading the authors to advocate for a taxonomy of knowledge expression in human-AI collaborative systems.

0 citationsRead paper

A Hybrid Intrusion Detection System for Electric Vehicle Charging Infrastructure

Jun 22, 2026

This study addresses the heightened security risks faced by electric vehicle (EV) charging infrastructure upon integration into smart grids, particularly the threat of cross-domain cyber-physical attacks that existing intrusion detection systems inadequately cover. To counter this challenge, the authors propose a novel two-tier collaborative hybrid intrusion detection system that synergistically integrates network-based (NIDS) and host-based (HIDS) approaches. This framework achieves, for the first time in EV charging scenarios, deep fusion of dual-source data to effectively detect sophisticated composite attacks spanning both cyber and physical layers. Experimental evaluation on the CICEVSE2024 dataset demonstrates that the NIDS component attains a 99.99% accuracy in identifying network-level attacks, while the HIDS component achieves an 83.47% accuracy in detecting false data injection attacks (FDIA), cryptojacking, backdoors, and most denial-of-service (DoS) and reconnaissance attacks—significantly outperforming standalone detection methods.

0 citationsRead paper

ICAN-Deploy: Identity-Stable Canary Deployment for Safety-Critical Embodied Agents

May 27, 2026

This work addresses the challenge of repeated re-certification in traditional canary deployments for safety-critical embodied intelligent agents, which arises from changes to the system’s cryptographic identity. The authors propose ICAN-Deploy, a middleware that decouples immutable capability names—hashed to serve as stable identities—from mutable capability versions, thereby preserving identity hashes throughout the canary window and enabling, for the first time, identity-stable canary deployment. The approach integrates a state machine design, a runtime governance layer, AST-based static analysis, closed-form proofs, and TLA+ model checking to guarantee safety and correctness. Empirical validation on a Franka Panda robotic arm in MuJoCo across 100 deployments demonstrates zero identity drift, with entry latency falling within a 95% BCa confidence interval of [1.52, 2.01] milliseconds.

0 citationsRead paper