mbrl-lib
mbrl-lib copied to clipboard
[Feature Request] (partially) known dynamics
Hi, Usually in practice we have some dynamic model which at least approximate the true environment x_{t+1} = f(x_t, u_t) (e.g from Newton laws). My question is how to include this a priori knowledge in model of dynamics? It seems like we could have better starting point, faster learning, stability and reasonable output
Regards,
Hi. Thanks for the interest in the library. This is definitely an interesting research direction, and I'm not sure there is a definite answer yet. This paper seems relevant. Is this more or less what you have in mind?
Thanks, this is exactly what I was asking, great survey.