发表机构
NC State University; Duke University（北卡罗来纳州立大学; 杜克大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本文提出ASCEND智能体，在研究者笔记本上运行，通过认证连接管理HPC集群和GPU工作站，实现故障恢复、结果复现、并行加速等自主科学计算，并验证了策略验证层的高效性。
AI 中文摘要
传统科学计算要求研究人员将计算意图转化为环境配置、资源请求和可执行作业，然后从调度器状态和应用日志中诊断故障。我们提出了ASCEND（自主科学计算引擎与新颖发现），一种由AI驱动的智能体接口，该智能体运行在研究人员自己的笔记本电脑上，通过多路复用的认证连接访问Slurm管理的集群和GPU工作站，并由本地执行的工具检查特定站点的执行策略；语言模型远程托管且不持有任何凭证。无需设施级服务：每个资源上的账户即可，公共安装程序允许用户链接自己的其他Slurm集群或工作站。我们报告了四个记录案例：（1）智能体在植入的张量设备故障上闭合了故障恢复循环，提交、诊断、修复并重新提交，同时保留了作业级工件；（2）它从作者发布的预测中复现了天气预报模型的已发表评估，与已发表曲线在z500上相差2.1%，在t850上相差2.4%，同时识别出论文文本中的单位差异和发布数据中的初始化字段差异；（3）它在逐位一致性的要求下并行化了一个已发布的12,693行地球物理求解器，将墙钟运行时间从约十二小时减少到约两小时；（4）该要求暴露了已发布求解器中的两个未定义行为实例，均已修复并上报上游。另外，对策略层的预先指定评估发现，部署的验证器拒绝了30个构造违规中的29个，并保留剩余一个以供批准，同时拒绝了14个合法请求中的3个。自主性在作者监督下行使；端到端恢复基准测试仍有待完成。
英文摘要
Traditional scientific computing requires researchers to translate intent into environment configuration, resource requests, and executable jobs, then diagnose failures from scheduler and application logs. We present ASCEND (Autonomous Scientific Computing Engine and Novel Discovery), an AI-powered agent system, operated through a command-line or a local web chat interface, that runs the agent on the researcher's own laptop, reaching Slurm clusters and a GPU workstation over multiplexed authenticated SSH, with site-specific execution policies checked by locally executed tools. The language-model agent (Claude Code or Codex, chosen at each launch) is hosted remotely and holds no credentials; an account on each resource suffices, and a public installer links additional clusters or workstations. We report four recorded cases: (1) the agent closed a failure-recovery loop on a planted tensor-device fault, submitting, diagnosing, repairing, and resubmitting it; (2) it reproduced the published evaluation of a weather-forecasting model, recovering an incompletely stated evaluation protocol and agreeing with the published curves to 2.1% (z500) and 2.4% (t850), while identifying two discrepancies in the paper's released materials; (3) it parallelized a released 12,693-line geophysical solver under a requirement of bit-for-bit identity of exported outputs, reducing runtime from about twelve hours to about two; (4) that requirement exposed two instances of undefined behavior in the published solver, both repaired and reported upstream. These cases used ordinary scheduler commands; a separate, pre-specified evaluation of the optional policy validator rejected 29 of 30 constructed violations, held one for approval, and denied 3 of 14 legitimate requests. Autonomy was exercised under author supervision, and agent transcripts were not retained, so intervention rates are not independently verifiable.
Comments21 pages, 8 figures, 6 tables. Code and installer: https://github.com/jpliu168/ASCEND
