API#

Interpreter is used to fit Decision tree policy to Neural networks policy.

Main class#

interpreter.Interpreter(oracle, learner, env)

A class to interpret a neural net policy using a decision tree policy.

interpreter.policies.ObliqueDTPolicy(clf, env)

Oblique Decision Tree Policy class.

Main functions#

interpreter.interpreter.Interpreter.fit(...)

Train the decision tree policy using data generated by the oracle.

interpreter.policies.ObliqueDTPolicy.get_oblique_data(S)

Generate oblique data by creating pairwise differences between observations.