Nesting Particle Filters for Experimental Design in Dynamical Systems
This work addresses experimental design in dynamical systems for researchers, but it is incremental as it builds on existing amortized techniques with a novel method.
The paper tackles Bayesian experimental design for non-exchangeable data by formulating it as risk-sensitive policy optimization, resulting in the Inside-Out SMC^2 algorithm that shows efficacy in numerical validation on dynamical systems compared to state-of-the-art strategies.
In this paper, we propose a novel approach to Bayesian experimental design for non-exchangeable data that formulates it as risk-sensitive policy optimization. We develop the Inside-Out SMC$^2$ algorithm, a nested sequential Monte Carlo technique to infer optimal designs, and embed it into a particle Markov chain Monte Carlo framework to perform gradient-based policy amortization. Our approach is distinct from other amortized experimental design techniques, as it does not rely on contrastive estimators. Numerical validation on a set of dynamical systems showcases the efficacy of our method in comparison to other state-of-the-art strategies.