Inthiswork,we1 propose thefirstfinite-time system identification algorithm forpartiallyobservable linear dynamical systems (LDS)2 inadaptive and closed-loop settings. Prior estimation methods only work when the actions/controls are iidrandom3 noise and do not allow for any exploitation or strategic exploration. Our proposed estimation5 algorithm allows the data collection with an adaptive controller and the design of fully adaptive RL methods. We6 believe this contribution alone has a great interest in both RL and control communities. Note that prior works in this area, such as [6,12,37,40-46] have been published in recent machine learning21 conferences(NeurIPS,ICML...).

This sampling process is embodied as the policyπ, which is responsible for outputting an actiona conditioned on past environmental states{s}.

We present an algorithm for learning switched linear dynamical systems in discrete time from noisy observations of the system's full state or output. Switched linear systems use multiple linear dynamical modes to fit the data within some desired tolerance. They arise quite naturally in applications to robotics and cyberphysical systems. Learning switched systems from data is a NP-hard problem that is nearly identical to the k-linear regression problem of fitting k > 1 linear models to the data. A direct mixed-integer linear programming (MILP) approach yields time complexity that is exponential in the number of data points. In this paper, we modify the problem formulation to yield an algorithm that is linear in the size of the data while remaining exponential in the number of state variables and the desired number of modes. To do so, we combine classic ideas from the ellipsoidal method for solving convex optimization problems, and well-known oracle separation results in non-smooth optimization. We demonstrate our approach on a set of microbenchmarks and a few interesting real-world problems. Our evaluation suggests that the benefits of this algorithm can be made practical even against highly optimized off-the-shelf MILP solvers.

Over time,ADAPTON improves its model estimates and obtains more accurate gradient updates to improve the controller.