Reuben

Thanks! This is a bit of a basic question, but what's the best way of using #190? That is, if I clone and build that PR, how do I then...

Ah, very cool, I think this will be useful! What do the a1, a2... represent here? I'm not quite sure I understand exactly what the function does.

I have a question about runtime and concurrency. I notice that the framerate of the simulation slows when I have a large number of particles. It would be nice if...

Btw, I refactored some of your code as follows, to give an alternate way of looking at it: ```haskell prior :: (MonadSample m, Diff td ~ Float) => BehaviourF m...

Nice! I've been writing a tutorial in a notebook, and trying a few variations of the model, which I'll quickly describe. Variations: - have a constant parameter be jointly inferred...

(Btw I started putting together an example in a notebook, and will extend it to a tutorial. Draft here: https://monad-bayes-site.netlify.app/realtimeinference)

Cool! I think I understand how dunai and Rhine's FRP work, and the implementation of the particle filter, but the details of how to combine clocks with schedules, how to...

OK, I have a Rhine question: is there a simple way to shift a signal forward or backwards? I'd like to do `sf >>> shift @Millisecond 300`. (I can see...

> Ah, how did that work out? Can I see the WIP code somewhere? I put my WIP at https://github.com/reubenharry/rhine-bayes-examples Here's where I track the past/future: https://github.com/reubenharry/rhine-bayes-examples/blob/master/src/Future.hs

Thanks for your advice! In the medium term, the first option seems appealing, particularly since we're also working in Jax. The kernel itself is actually quite simple; the difficulty is...