Concepts for Securing Agentic AI Coding and the Terok Environment
本文评估了Agentic AI在软件开发中的IT安全风险，并提出了一种缓解这些风险的概念及其实现方案，旨在促进其安全应用。
0 citationsRead paper
Selected work
Representative Papers
Optimizations on Graph-Level for Domain Specific Computations in Julia and Application to QED
To address the challenges of scheduling heterogeneous subtasks and low hardware utilization in scientific computing, this paper proposes a domain-semantic-aware DAG-driven scheduling and compilation framework. Methodologically, it models computational workflows as directed acyclic graphs (DAGs) and incorporates domain-specific physical constraints—such as those from quantum electrodynamics—to jointly optimize cross-device parallelism, data movement, and dependency management. Integrating static compilation with dynamic scheduling, the framework enables fine-grained resource allocation and automatic code generation within Julia. Its key contribution lies in being the first to deeply embed domain-specific semantics throughout the entire DAG scheduling and compilation pipeline, thereby overcoming the limitations of conventional hardware-agnostic schedulers. Experimental evaluation on multi-external-particle scattering matrix element computation demonstrates significant improvements in execution efficiency and scalability, with hardware utilization increased by up to 2.3×.
0 citationsRead paper
Machine Learning Time Propagators for Time-Dependent Density Functional Theory Simulations
This work addresses the high computational cost of electron dynamics simulations in time-dependent density functional theory (TDDFT). We propose an autoregressive neural operator as a time propagator to efficiently model the evolution of electron density under time-varying external fields—such as laser pulses—while preserving physical fidelity. Our method integrates physics-informed constraints (e.g., continuity equation, energy conservation priors) with multi-scale feature engineering to construct a high-resolution, real-space sequence prediction model. Evaluated on a class of one-dimensional diatomic molecular systems, the model achieves TDDFT-level accuracy while accelerating inference by one to two orders of magnitude compared to conventional numerical solvers. It enables real-time parameter tuning and long-time-scale dynamical simulations. This work establishes a scalable, physics-aware machine learning paradigm for modeling ultrafast electronic responses of complex materials under intense laser irradiation.
0 citationsRead paper
Data Version Management and Machine-Actionable Reproducibility for HPC based on git and DataLad
This work addresses two critical limitations of Git-based data versioning tools (e.g., DataLad) in HPC environments: incompatibility with the Slurm batch scheduler and poor I/O efficiency on parallel file systems (e.g., Lustre, GPFS). We propose the first lightweight, non-intrusive framework that deeply integrates DataLad with Slurm. Our approach extends Slurm’s job encapsulation mechanism and introduces automated, fine-grained metadata capture at the job level, thereby ensuring end-to-end reproducibility. Furthermore, we optimize versioning operations—such as dataset checkout and commit—by adapting their I/O paths for parallel file systems. Evaluation on a production supercomputing cluster demonstrates a 92% reduction in metadata capture overhead and a 3.8× speedup in large-dataset version switching. These improvements significantly broaden the applicability of data version control to production-scale HPC batch workflows.
0 citationsRead paper
Recent publications
Latest Papers
Concepts for Securing Agentic AI Coding and the Terok Environment
本文评估了Agentic AI在软件开发中的IT安全风险，并提出了一种缓解这些风险的概念及其实现方案，旨在促进其安全应用。
0 citationsRead paper
Optimizations on Graph-Level for Domain Specific Computations in Julia and Application to QED
To address the challenges of scheduling heterogeneous subtasks and low hardware utilization in scientific computing, this paper proposes a domain-semantic-aware DAG-driven scheduling and compilation framework. Methodologically, it models computational workflows as directed acyclic graphs (DAGs) and incorporates domain-specific physical constraints—such as those from quantum electrodynamics—to jointly optimize cross-device parallelism, data movement, and dependency management. Integrating static compilation with dynamic scheduling, the framework enables fine-grained resource allocation and automatic code generation within Julia. Its key contribution lies in being the first to deeply embed domain-specific semantics throughout the entire DAG scheduling and compilation pipeline, thereby overcoming the limitations of conventional hardware-agnostic schedulers. Experimental evaluation on multi-external-particle scattering matrix element computation demonstrates significant improvements in execution efficiency and scalability, with hardware utilization increased by up to 2.3×.
0 citationsRead paper
Machine Learning Time Propagators for Time-Dependent Density Functional Theory Simulations
This work addresses the high computational cost of electron dynamics simulations in time-dependent density functional theory (TDDFT). We propose an autoregressive neural operator as a time propagator to efficiently model the evolution of electron density under time-varying external fields—such as laser pulses—while preserving physical fidelity. Our method integrates physics-informed constraints (e.g., continuity equation, energy conservation priors) with multi-scale feature engineering to construct a high-resolution, real-space sequence prediction model. Evaluated on a class of one-dimensional diatomic molecular systems, the model achieves TDDFT-level accuracy while accelerating inference by one to two orders of magnitude compared to conventional numerical solvers. It enables real-time parameter tuning and long-time-scale dynamical simulations. This work establishes a scalable, physics-aware machine learning paradigm for modeling ultrafast electronic responses of complex materials under intense laser irradiation.
0 citationsRead paper
Data Version Management and Machine-Actionable Reproducibility for HPC based on git and DataLad
This work addresses two critical limitations of Git-based data versioning tools (e.g., DataLad) in HPC environments: incompatibility with the Slurm batch scheduler and poor I/O efficiency on parallel file systems (e.g., Lustre, GPFS). We propose the first lightweight, non-intrusive framework that deeply integrates DataLad with Slurm. Our approach extends Slurm’s job encapsulation mechanism and introduces automated, fine-grained metadata capture at the job level, thereby ensuring end-to-end reproducibility. Furthermore, we optimize versioning operations—such as dataset checkout and commit—by adapting their I/O paths for parallel file systems. Evaluation on a production supercomputing cluster demonstrates a 92% reduction in metadata capture overhead and a 3.8× speedup in large-dataset version switching. These improvements significantly broaden the applicability of data version control to production-scale HPC batch workflows.
0 citationsRead paper