Selected work

Representative Papers

Parameter-Dependent LMI Synthesis for Semi-Global Differential ISS Trajectory Tracking of Nonholonomic Mobile Robots Under Multiplicative Wheel Slip

Aug 08, 2026

This work addresses trajectory tracking for nonholonomic mobile robots operating on highly variable terrain where severe multiplicative wheel slip significantly degrades performance. The authors propose a control approach based on parameter-dependent linear matrix inequalities (LMIs), which explicitly constructs an additive disturbance upper bound for multiplicative wheel slip within the Kanayama error coordinates—a first in the literature. By integrating affine parameterized gains with a nonlinear storage function, the method synergistically combines physical modeling insights with convex optimization. It further incorporates sampled convexification, variational contraction analysis, forward invariance, and dissipativity-based trajectory reconstruction to guarantee semi-global differential input-to-state stability with a prescribed exponential decay rate. Experimental results demonstrate that, over complex terrain featuring six segments with bidirectional ±50% slip, the proposed controller reduces peak tracking error by 12% and 49% compared to fixed-gain LMI and hand-tuned baselines, respectively, while satisfying certified envelope constraints in all 100 Monte Carlo trials.

0 citationsRead paper

HugSelect: An Explainable Multi-Criteria Decision-Support Framework for foundation-model selection

Aug 08, 2026

This work addresses the lack of systematic, auditable decision support in foundational model selection, a process often driven by popularity rather than functional capabilities, operational constraints, or community-based quality assessments. To remedy this, we propose HugSelect—the first interpretable multi-criteria decision framework that formalizes foundational model selection as a traceable software component selection task. HugSelect integrates model metadata, functional attributes, and community-perceived quality into a unified knowledge base and a decomposable weighted scoring system. Evaluated on 71,274 models, our approach achieves an F1 score of 0.801 in functional feature extraction and 0.84 accuracy in quality mapping. It attains model-level Coverage@10 of 0.61 and family-level Coverage@10 of 0.91, matching the recommendation performance of leading commercial systems while offering transparent, interpretable results that elicit positive user feedback.

0 citationsRead paper

A Novel Gravity-Quasi-Laplacian Approach to Identifying Influential Nodes in Complex Networks

Jul 25, 2026

Existing methods for identifying influential nodes often suffer from limited accuracy, poor resolution, reliance on tunable parameters, and high computational complexity. This work proposes a parameter-free, interpretable, and efficient gravitational quasi-Laplacian framework that constructs structural representations using node degree and k-shell index, incorporates a quasi-Laplacian operator to capture local topology, and employs a short-range gravitational aggregation mechanism with a fixed radius (R=3) to evaluate node influence. Requiring no parameter tuning, the proposed method significantly outperforms eight state-of-the-art algorithms across nine real-world networks, demonstrating superior performance in identification accuracy, resolution, and computational efficiency.

0 citationsRead paper

TLRNet: Estimating Individual Treatment Effect based on Local Information and Single Learner Structure

Jul 23, 2026

This work addresses the problem of individualized treatment effect (ITE) estimation by proposing a novel pseudo-single learner architecture based on deep neural networks. The method employs a single estimator to jointly model potential outcomes under both treatment conditions and enhances causal representation learning through local information augmentation, thereby substantially simplifying model architecture while improving estimation efficiency. Experimental results on the IHDP benchmark dataset demonstrate that the proposed approach achieves state-of-the-art performance in ITE estimation accuracy, confirming its effectiveness and practical feasibility.

0 citationsRead paper

Uncertainty-Aware Fuzzy Centrality Measures for Influential Node Identification: A Structural Modeling Approach Toward E-Commerce Applications

Apr 26, 2026

This study addresses the challenge of identifying influential nodes in e-commerce platforms, where user–item interactions are inherently uncertain and noisy, rendering traditional deterministic network models inadequate. To overcome this limitation, the work proposes a novel fuzzy centrality measure that integrates fuzzy graph theory with uncertainty modeling. The method captures implicit interaction relationships through structural modeling of fuzzy connections and incorporates structural embedding techniques to construct a robust influence assessment metric suitable for noisy environments. Experimental evaluation on real-world e-commerce datasets demonstrates that the proposed approach significantly outperforms state-of-the-art centrality algorithms in both accuracy and robustness for influential node identification.

0 citationsRead paper

Recent publications

Latest Papers

Parameter-Dependent LMI Synthesis for Semi-Global Differential ISS Trajectory Tracking of Nonholonomic Mobile Robots Under Multiplicative Wheel Slip

Aug 08, 2026

This work addresses trajectory tracking for nonholonomic mobile robots operating on highly variable terrain where severe multiplicative wheel slip significantly degrades performance. The authors propose a control approach based on parameter-dependent linear matrix inequalities (LMIs), which explicitly constructs an additive disturbance upper bound for multiplicative wheel slip within the Kanayama error coordinates—a first in the literature. By integrating affine parameterized gains with a nonlinear storage function, the method synergistically combines physical modeling insights with convex optimization. It further incorporates sampled convexification, variational contraction analysis, forward invariance, and dissipativity-based trajectory reconstruction to guarantee semi-global differential input-to-state stability with a prescribed exponential decay rate. Experimental results demonstrate that, over complex terrain featuring six segments with bidirectional ±50% slip, the proposed controller reduces peak tracking error by 12% and 49% compared to fixed-gain LMI and hand-tuned baselines, respectively, while satisfying certified envelope constraints in all 100 Monte Carlo trials.

0 citationsRead paper

HugSelect: An Explainable Multi-Criteria Decision-Support Framework for foundation-model selection

Aug 08, 2026

This work addresses the lack of systematic, auditable decision support in foundational model selection, a process often driven by popularity rather than functional capabilities, operational constraints, or community-based quality assessments. To remedy this, we propose HugSelect—the first interpretable multi-criteria decision framework that formalizes foundational model selection as a traceable software component selection task. HugSelect integrates model metadata, functional attributes, and community-perceived quality into a unified knowledge base and a decomposable weighted scoring system. Evaluated on 71,274 models, our approach achieves an F1 score of 0.801 in functional feature extraction and 0.84 accuracy in quality mapping. It attains model-level Coverage@10 of 0.61 and family-level Coverage@10 of 0.91, matching the recommendation performance of leading commercial systems while offering transparent, interpretable results that elicit positive user feedback.

0 citationsRead paper

A Novel Gravity-Quasi-Laplacian Approach to Identifying Influential Nodes in Complex Networks

Jul 25, 2026

Existing methods for identifying influential nodes often suffer from limited accuracy, poor resolution, reliance on tunable parameters, and high computational complexity. This work proposes a parameter-free, interpretable, and efficient gravitational quasi-Laplacian framework that constructs structural representations using node degree and k-shell index, incorporates a quasi-Laplacian operator to capture local topology, and employs a short-range gravitational aggregation mechanism with a fixed radius (R=3) to evaluate node influence. Requiring no parameter tuning, the proposed method significantly outperforms eight state-of-the-art algorithms across nine real-world networks, demonstrating superior performance in identification accuracy, resolution, and computational efficiency.

0 citationsRead paper

TLRNet: Estimating Individual Treatment Effect based on Local Information and Single Learner Structure

Jul 23, 2026

This work addresses the problem of individualized treatment effect (ITE) estimation by proposing a novel pseudo-single learner architecture based on deep neural networks. The method employs a single estimator to jointly model potential outcomes under both treatment conditions and enhances causal representation learning through local information augmentation, thereby substantially simplifying model architecture while improving estimation efficiency. Experimental results on the IHDP benchmark dataset demonstrate that the proposed approach achieves state-of-the-art performance in ITE estimation accuracy, confirming its effectiveness and practical feasibility.

0 citationsRead paper

Uncertainty-Aware Fuzzy Centrality Measures for Influential Node Identification: A Structural Modeling Approach Toward E-Commerce Applications

Apr 26, 2026

This study addresses the challenge of identifying influential nodes in e-commerce platforms, where user–item interactions are inherently uncertain and noisy, rendering traditional deterministic network models inadequate. To overcome this limitation, the work proposes a novel fuzzy centrality measure that integrates fuzzy graph theory with uncertainty modeling. The method captures implicit interaction relationships through structural modeling of fuzzy connections and incorporates structural embedding techniques to construct a robust influence assessment metric suitable for noisy environments. Experimental evaluation on real-world e-commerce datasets demonstrates that the proposed approach significantly outperforms state-of-the-art centrality algorithms in both accuracy and robustness for influential node identification.

0 citationsRead paper