feedforward-closedloop-learning 2.2.1

Feedforward Closedloop Learning (FCL)

Feedforward closed loop learning (FCL) is a learning algorithm which adds flexibility to autonomous agents.

A designer defines an initial behaviour as a reflex and then FCL learns from the reflex to develop new flexible behaviours.

The Python documentation can be obtained with:

import feedforward_closedloop_learning as fcl
help(fcl)

The Python API is identical to the C++ API: The documentation can be found here: https://glasgowneuro.github.io/feedforward_closedloop_learning/

The best way to get started is to look at the script in tests_py: https://github.com/glasgowneuro/feedforward_closedloop_learning/tree/master/tests_py

A full application using the Python API is our vizdoom agent: https://github.com/glasgowneuro/fcl_demos