Selected work

Representative Papers

Neural Variability Enhances Artificial Network Robustness

Jun 11, 2026

This work addresses the limited robustness of artificial neural networks under adversarial attacks and natural image corruptions, a challenge often exacerbated by the neglect of structured noise in neural activations. The authors propose a biologically inspired local noise mechanism that models structured noise by analyzing the covariance structure of activations induced by clean and perturbed inputs. Relying solely on local information, this approach is the first to systematically reveal how structured noise differentially enhances robustness across perturbation types. Experimental results demonstrate that the proposed strategy significantly improves model robustness against natural corruptions, and notably, the noise structures learned under adversarial attacks exhibit strong generalization to other attack variants.

0 citationsRead paper

The Association of Transformer-based Sentiment Analysis with Symptom Distress and Deterioration in Routine Psychotherapy Care

May 10, 2026

This study leverages natural language processing to quantify emotional dynamics in psychotherapy and examine their association with clinical distress and risk of deterioration. Building upon a Transformer architecture, the authors develop a fine-grained sentiment analysis model to extract utterance- and session-level emotional features from 751 therapy dialogues, which are then statistically linked to OQ-45 scale scores. For the first time, Transformer-derived emotional features are treated as standalone psychometric indicators, revealing significant correlations with the emotion-related subscales of the OQ-45. Moreover, these features exhibit marked differences in patients at high risk of symptom deterioration or premature termination, thereby extending the applicability of artificial intelligence in psychological assessment.

0 citationsRead paper

PRIMRose: Insights into the Per-Residue Energy Metrics of Proteins with Double InDel Mutations using Deep Learning

Dec 06, 2025

This study addresses the challenge of predicting the structural and functional impacts of double amino acid insertion/deletion (InDel) mutations. We propose the first residue-level local energy perturbation prediction method, leveraging Rosetta-computed multidimensional energy features. A convolutional neural network is trained and validated on a large-scale, multi-source dataset comprising nearly 300,000 double InDel variants. The model achieves high accuracy in predicting multiple energy-based metrics—including van der Waals, solvation, and hydrogen-bonding energies—and identifies solvent accessibility and secondary structure context as key determinants of mutational tolerance. Furthermore, it successfully pinpoints residue-level mutational hotspots with elevated tolerance. This work introduces the first interpretable, residue-resolution energy analysis tool for double InDels, enabling mechanistic insights into pathogenic variants and facilitating rational protein engineering design.

0 citationsRead paper

The Impact of Trade and Financial Openness on Operational Efficiency and Growth: Evidence from Turkish Banks

Dec 02, 2025

This study examines how trade and financial openness affected operational efficiency and growth of Turkish banks during 2010–2023. Using a CAMELG-DEA framework to measure bank efficiency and dynamic panel GMM estimation, it integrates macro-level openness indicators with micro-level bank data. Results show that trade openness enhances operational efficiency primarily through expanded international banking activities, whereas financial openness stimulates credit expansion and non-interest income growth—but its impact is dampened by domestic poverty levels. The study provides the first empirical evidence from a developing economy distinguishing the heterogeneous transmission channels through which trade versus financial openness affect bank performance. It further identifies domestic institutional conditions—particularly poverty—as critical moderators of openness-related gains, thereby offering micro-level evidence to inform sequencing and complementary policy design for financial liberalization.

0 citationsRead paper

Post Processing Graphical User Interface for Heat Flow Visualization

Nov 11, 2025

Thermal control engineers lack efficient tools for extracting and visualizing thermal flow data from Thermal Desktop (TD), resulting in inefficient post-processing. To address this, we propose a MATLAB/C++ hybrid GUI system that integrates the OpenTD API with a custom CSR file parser. Leveraging an implicit node–path–submodel ID mapping embedded in CSR files—exploited via a “side-effect” mechanism—the system enables millisecond-level association and loading of thermal flows, temperatures, admittances, and submodel metrics. This approach improves data-matching efficiency by two to three orders of magnitude, substantially reducing post-processing time. The system bridges a critical gap in the TD ecosystem by enabling deep, interactive visualization of thermal flow metrics. Moreover, it provides a reproducible technical pathway and empirical foundation for enhancing thermal analysis capabilities in future OpenTD releases.

0 citationsRead paper

Recent publications

Latest Papers

Neural Variability Enhances Artificial Network Robustness

Jun 11, 2026

This work addresses the limited robustness of artificial neural networks under adversarial attacks and natural image corruptions, a challenge often exacerbated by the neglect of structured noise in neural activations. The authors propose a biologically inspired local noise mechanism that models structured noise by analyzing the covariance structure of activations induced by clean and perturbed inputs. Relying solely on local information, this approach is the first to systematically reveal how structured noise differentially enhances robustness across perturbation types. Experimental results demonstrate that the proposed strategy significantly improves model robustness against natural corruptions, and notably, the noise structures learned under adversarial attacks exhibit strong generalization to other attack variants.

0 citationsRead paper

The Association of Transformer-based Sentiment Analysis with Symptom Distress and Deterioration in Routine Psychotherapy Care

May 10, 2026

This study leverages natural language processing to quantify emotional dynamics in psychotherapy and examine their association with clinical distress and risk of deterioration. Building upon a Transformer architecture, the authors develop a fine-grained sentiment analysis model to extract utterance- and session-level emotional features from 751 therapy dialogues, which are then statistically linked to OQ-45 scale scores. For the first time, Transformer-derived emotional features are treated as standalone psychometric indicators, revealing significant correlations with the emotion-related subscales of the OQ-45. Moreover, these features exhibit marked differences in patients at high risk of symptom deterioration or premature termination, thereby extending the applicability of artificial intelligence in psychological assessment.

0 citationsRead paper

PRIMRose: Insights into the Per-Residue Energy Metrics of Proteins with Double InDel Mutations using Deep Learning

Dec 06, 2025

This study addresses the challenge of predicting the structural and functional impacts of double amino acid insertion/deletion (InDel) mutations. We propose the first residue-level local energy perturbation prediction method, leveraging Rosetta-computed multidimensional energy features. A convolutional neural network is trained and validated on a large-scale, multi-source dataset comprising nearly 300,000 double InDel variants. The model achieves high accuracy in predicting multiple energy-based metrics—including van der Waals, solvation, and hydrogen-bonding energies—and identifies solvent accessibility and secondary structure context as key determinants of mutational tolerance. Furthermore, it successfully pinpoints residue-level mutational hotspots with elevated tolerance. This work introduces the first interpretable, residue-resolution energy analysis tool for double InDels, enabling mechanistic insights into pathogenic variants and facilitating rational protein engineering design.

0 citationsRead paper

The Impact of Trade and Financial Openness on Operational Efficiency and Growth: Evidence from Turkish Banks

Dec 02, 2025

This study examines how trade and financial openness affected operational efficiency and growth of Turkish banks during 2010–2023. Using a CAMELG-DEA framework to measure bank efficiency and dynamic panel GMM estimation, it integrates macro-level openness indicators with micro-level bank data. Results show that trade openness enhances operational efficiency primarily through expanded international banking activities, whereas financial openness stimulates credit expansion and non-interest income growth—but its impact is dampened by domestic poverty levels. The study provides the first empirical evidence from a developing economy distinguishing the heterogeneous transmission channels through which trade versus financial openness affect bank performance. It further identifies domestic institutional conditions—particularly poverty—as critical moderators of openness-related gains, thereby offering micro-level evidence to inform sequencing and complementary policy design for financial liberalization.

0 citationsRead paper

Post Processing Graphical User Interface for Heat Flow Visualization

Nov 11, 2025

Thermal control engineers lack efficient tools for extracting and visualizing thermal flow data from Thermal Desktop (TD), resulting in inefficient post-processing. To address this, we propose a MATLAB/C++ hybrid GUI system that integrates the OpenTD API with a custom CSR file parser. Leveraging an implicit node–path–submodel ID mapping embedded in CSR files—exploited via a “side-effect” mechanism—the system enables millisecond-level association and loading of thermal flows, temperatures, admittances, and submodel metrics. This approach improves data-matching efficiency by two to three orders of magnitude, substantially reducing post-processing time. The system bridges a critical gap in the TD ecosystem by enabling deep, interactive visualization of thermal flow metrics. Moreover, it provides a reproducible technical pathway and empirical foundation for enhancing thermal analysis capabilities in future OpenTD releases.

0 citationsRead paper