This is code for iDAD [1] modified for use in the paper "Nesting Particle Filters for Experimental Design in Dynamical Systems" [2].
To set up the Python environment and run experiments, follow the instructions at the original repo. The four experiments implemented for [2] are pendulum_linear.py
, pendulum.py
, cartpole.py
, and double_pendulum.py
. The sPCE lower bound for the trained models can be computed using eval_sPCE.py
. To compute the EIG estimate for the trained models, run the respective file {experiment_name}_eig.py
.
[1] Ivanova, D. R., Foster, A., Kleinegesse, S., Gutmann, M., and Rainforth, T. Implicit deep adaptive design: Policy–based experimental design without likelihoods. In Advances in Neural Information Processing Systems. 2021. Paper. Code.
[2] Iqbal, S., Corenflos, A., Särkkä, S., and Abdulsamad, H. Nesting Particle Filters for Experimental Design in Dynamical Systems. In International Conference on Machine Learning. 2024. Paper. Code.