New zero-one law simplifies one-shot system identification

Researchers have developed a new method for identifying analytic systems from a single experiment, applicable to systems linearly parameterized by prescribed dictionary terms. They proved a sharp zero-one law, indicating that either no input uniquely determines the coefficients or almost every random input from a non-degenerate Gaussian measure does. This approach simplifies one-shot system identification and provides a certificate for recovered models, with numerical examples demonstrating its effectiveness in recovering dynamical systems, nonlinear partial differential equations, and structured matrix families from single trajectory data. AI

IMPACT This research advances theoretical understanding in system identification, potentially impacting AI model training and validation.

RANK_REASON The item is an academic paper submitted to arXiv on numerical analysis. [lever_c_demoted from research: ic=1 ai=0.7]

Read on arXiv cs.LG →

• alphaXiv
• arXiv
• A zero-one law for one-shot system identification
• CatalyzeX Code Finder for Papers
• CORE Recommender
• DagsHub
• Gaussian measure
• Gotit.pub
• Hugging Face
• Influence Flower
• numerical analysis
• ScienceCast

• paper
• other

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