The authors are currently working on the next iteration of the model which will hopefully come with a few improvements.

AFAIK, they are keen on getting their hands on data that proved difficult for the model -- I suggest you get in touch with them if you think you could help!