Research map / Learning theory and algorithms / Sample complexity

Robust System Identification: Finite-sample Guarantees and Connection to Regularization

ICLR 2025 · Learning theory and algorithms › Sample complexity · approach: Sample complexity of learning distributions

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes a robust least-squares estimation framework for learning nonlinear dynamical systems from a single trajectory, with finite-sample guarantees and a connection to regularization.

Problem

Learning nonlinear dynamical systems from a single sample trajectory, where the least squares estimate has poor identification error when the sample size is small or the model misspecifies the true dynamics.

Approach

Incorporates robust optimization into least squares estimation for system identification, yielding a robust LSE whose finite-sample behavior is analyzed and shown to connect to regularization.

Experimental setup

Baselines: least squares estimate (LSE)

Builds on: least squares estimation

Closest papers by meaning

• Finite Sample Analyses for Continuous-time Linear Systems: System Identification and Online Control (NeurIPS 2025 · similarity 0.88)
• Identification of Analytic Nonlinear Dynamical Systems with Non-asymptotic Guarantees (NeurIPS 2024 · similarity 0.88)
• Near optimal finite time identification of arbitrary linear dynamical systems (ICML 2019 · similarity 0.88)
• Learning the Uncertainty Sets of Linear Control Systems via Set Membership: A Non-asymptotic Analysis (ICML 2024 · similarity 0.86)
• Robust-Adaptive Control of Linear Systems: beyond Quadratic Costs (NeurIPS 2020 · similarity 0.86)
• Safely Learning Controlled Stochastic Dynamics (NeurIPS 2025 · similarity 0.84)