New CENTILE model streamlines telemetry analysis for HPC and networking

Researchers have developed CENTILE, a new generative foundation model designed for network and systems telemetry. Unlike previous approaches that used separate predictors for each task, CENTILE is pretrained once on an operator's event streams to handle heterogeneous telemetry as event-driven, irregularly timed entity streams. This model is evaluated by replaying decisions driven by its calibrated conditional quantiles, aiming to improve operational outcomes. Experiments show CENTILE significantly enhances High-Performance Computing (HPC) scheduling and network provisioning decisions, reducing slowdowns and violation rates. AI

IMPACT This model could significantly improve the efficiency and decision-making processes in complex computing and networking environments.

RANK_REASON The cluster describes a new research paper detailing a novel foundation model for telemetry analysis. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• arXiv
• CatalyzeX
• Connected Papers
• DagsHub
• Gotit.pub
• high-performance computing
• Hugging Face
• Litmaps
• ScienceCast
• scite Smart Citations
• telemetry

• paper
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →