Selected work

Representative Papers

The Hamilton-Jacobi Theory of Deep Learning

May 27, 2026

This work establishes a precise mathematical correspondence between deep neural network training and inference and the theory of partial differential equations, interpreting training as solving an initial-value problem for the Hamilton–Jacobi equation. Each gradient update step is equivalent to selecting an initial condition for a viscous Hamilton–Jacobi equation and optimally fitting data via the Hopf–Cole propagator, while inference corresponds to evaluating the solution at specific points. By introducing a single deformation parameter ε, the framework unifies four perspectives—Hamilton–Jacobi PDEs, tropical geometry, convex optimization, and network architecture—for the first time. It yields a minimax optimal generalization rate of O(n⁻¹⁄⁽ᵈ⁺²⁾) at fixed time, reveals how ε governs adversarial robustness, provides an O(N) closed-form influence function, and characterizes the fold bifurcation of the entropy landscape induced by varying ε.

0 citationsRead paper

Beyond Model Readiness: Institutional Readiness for AI Deployment in Public Systems

May 16, 2026

This study addresses the persistent challenge of scaling artificial intelligence systems in the public sector, where institutional barriers—such as inadequate approval processes, weak data governance, insufficient oversight capacity, fiscal unsustainability, or regulatory ambiguity—often impede deployment despite technical feasibility. To bridge this gap, the authors propose the Institutional Alignment Readiness (IAR) framework, which shifts focus from the AI model itself to the receiving institution. The framework assesses readiness across five dimensions: institutional and operational compatibility, data ecosystem maturity, human oversight capacity, fiscal sustainability, and regulatory alignment, offering a practical evaluation tool tailored for resource-constrained settings. Through qualitative analysis of two anonymized public education system cases, the IAR framework effectively identifies institutional bottlenecks and informs phased deployment decisions—ranging from prohibition and piloting to full-scale implementation—thereby addressing a critical void in existing AI assessment methodologies at the institutional level.

0 citationsRead paper

Are LLMs reliable? An exploration of the reliability of large language models in clinical note generation

May 21, 2025

This study addresses the reliability bottleneck in large language models (LLMs) for clinical note generation (CNG), stemming from response variability. We propose the first multidimensional reliability evaluation framework tailored to healthcare settings, assessing string-level consistency, semantic consistency, and semantic correctness. We systematically evaluate 12 open- and closed-source LLMs using repeated sampling, prompt-consistency testing, and automated metrics (BLEU, BERTScore, natural language inference). All evaluations are validated by clinical domain experts. Results show that all models achieve >92% semantic consistency; Llama-70B attains the highest overall reliability—significantly outperforming most commercial LLMs. Moreover, smaller-parameter open-source models demonstrate superior accuracy, robustness, and compliance with local deployment and privacy requirements. This work establishes a reproducible, interpretable, and privacy-preserving evaluation paradigm for clinical deployment of LLM-driven CNG systems.

0 citationsRead paper

Recent publications

Latest Papers

The Hamilton-Jacobi Theory of Deep Learning

May 27, 2026

This work establishes a precise mathematical correspondence between deep neural network training and inference and the theory of partial differential equations, interpreting training as solving an initial-value problem for the Hamilton–Jacobi equation. Each gradient update step is equivalent to selecting an initial condition for a viscous Hamilton–Jacobi equation and optimally fitting data via the Hopf–Cole propagator, while inference corresponds to evaluating the solution at specific points. By introducing a single deformation parameter ε, the framework unifies four perspectives—Hamilton–Jacobi PDEs, tropical geometry, convex optimization, and network architecture—for the first time. It yields a minimax optimal generalization rate of O(n⁻¹⁄⁽ᵈ⁺²⁾) at fixed time, reveals how ε governs adversarial robustness, provides an O(N) closed-form influence function, and characterizes the fold bifurcation of the entropy landscape induced by varying ε.

0 citationsRead paper

Beyond Model Readiness: Institutional Readiness for AI Deployment in Public Systems

May 16, 2026

This study addresses the persistent challenge of scaling artificial intelligence systems in the public sector, where institutional barriers—such as inadequate approval processes, weak data governance, insufficient oversight capacity, fiscal unsustainability, or regulatory ambiguity—often impede deployment despite technical feasibility. To bridge this gap, the authors propose the Institutional Alignment Readiness (IAR) framework, which shifts focus from the AI model itself to the receiving institution. The framework assesses readiness across five dimensions: institutional and operational compatibility, data ecosystem maturity, human oversight capacity, fiscal sustainability, and regulatory alignment, offering a practical evaluation tool tailored for resource-constrained settings. Through qualitative analysis of two anonymized public education system cases, the IAR framework effectively identifies institutional bottlenecks and informs phased deployment decisions—ranging from prohibition and piloting to full-scale implementation—thereby addressing a critical void in existing AI assessment methodologies at the institutional level.

0 citationsRead paper

Are LLMs reliable? An exploration of the reliability of large language models in clinical note generation

May 21, 2025

This study addresses the reliability bottleneck in large language models (LLMs) for clinical note generation (CNG), stemming from response variability. We propose the first multidimensional reliability evaluation framework tailored to healthcare settings, assessing string-level consistency, semantic consistency, and semantic correctness. We systematically evaluate 12 open- and closed-source LLMs using repeated sampling, prompt-consistency testing, and automated metrics (BLEU, BERTScore, natural language inference). All evaluations are validated by clinical domain experts. Results show that all models achieve >92% semantic consistency; Llama-70B attains the highest overall reliability—significantly outperforming most commercial LLMs. Moreover, smaller-parameter open-source models demonstrate superior accuracy, robustness, and compliance with local deployment and privacy requirements. This work establishes a reproducible, interpretable, and privacy-preserving evaluation paradigm for clinical deployment of LLM-driven CNG systems.

0 citationsRead paper