Absolutely phenomenal! The work you are doing is amazing. Smooth presentations, readable understandable papers, and open source code. Thank you so much your work is closely related to my thesis. Keep up the good work, I wish you all the best.
We have a GitHub repo (and example) at github.com/patrick-kidger/NeuralCDE you can check out. We didn't specifically consider a forecasting problem in the paper but it should be straightforward to set up, using the classification example as a guide.
@@patrickkidger1285 nice, ill give a try, im new to neural ode and im currently learning how to solve/train in julia, checking the benchmark it should be also faster than pytorch, so i wonder why cde and other models like latent ode aren't done in julia. sorry my english.
Absolutely phenomenal!
The work you are doing is amazing.
Smooth presentations, readable understandable papers, and open source code.
Thank you so much your work is closely related to my thesis.
Keep up the good work, I wish you all the best.
Very interesting! And thanks for publishing the code, sadly not the standard for ML research.
Thankyou! I think publishing code is becoming more frequent these days, thankfully.
This is awesome! Could you please provide any insight on how to include estimates of time-varying parameters?
Nice talk! BTW, where can I download the slides? Thx!
A version of these slides are available here: kidger.site/links/CZszzuRmMmzo
could you please release a model/.py for forecasting?
We have a GitHub repo (and example) at github.com/patrick-kidger/NeuralCDE you can check out.
We didn't specifically consider a forecasting problem in the paper but it should be straightforward to set up, using the classification example as a guide.
@@patrickkidger1285 nice, ill give a try, im new to neural ode and im currently learning how to solve/train in julia, checking the benchmark it should be also faster than pytorch, so i wonder why cde and other models like latent ode aren't done in julia. sorry my english.