infer-actively / pymdp

A Python implementation of active inference for Markov Decision Processes
MIT License
450 stars 89 forks source link

Allow learning of likelihood and transition likelihood parameters with sequences of hidden state posteriors #109

Open conorheins opened 1 year ago

conorheins commented 1 year ago

Add functions to learning module (and update corresponding update_A() and update_B() methods of Agent class) to allow learning to incorporate an entire past sequence of hidden state posteriors, rather than just a single hidden state posterior.