Discrete-Time Sensitivity Analysis
An analytical tool in control and dynamical systems theory used to quantify the response of system outputs to small changes in parameters or initial conditions. By computing partial derivatives of trajectories with respect to parameters and recursing along time steps, it can reveal whether trajectory error converges or diverges when learned parameters deviate from the training optimum, and is an important mathematical method for analyzing closed-loop system stability.
Timeline (last 90 days)
Through discrete-time sensitivity analysis, the paper proves that parameters learned offline may be locally optimal near the training data, but when deployed in a closed-loop system, these parameters will precisely guarantee trajectory divergence.
Discrete-time sensitivity analysis is a tool in control and dynamical systems theory used to quantify the degree to which a system output responds to small changes in parameters or initial conditions, and its core is to compute the partial derivatives of a trajectory with respect to parameters and propagate them recursively along time steps.
All Facts (2)
Through discrete-time sensitivity analysis, the paper proves that parameters learned offline may be locally optimal near the training data, but when deployed in a closed-loop system, these parameters will precisely guarantee trajectory divergence.
50%UnverifiedDiscrete-time sensitivity analysis is a tool in control and dynamical systems theory used to quantify the degree to which a system output responds to small changes in parameters or initial conditions, and its core is to compute the partial derivatives of a trajectory with respect to parameters and propagate them recursively along time steps.
50%