3.10. SequentialLearning.AlgorithmDensity
Algorithm density
We define a density function that allows obtaining the law of the history under one algorithm from the law of the history under another algorithm when they are interacting with the same environment. This also requires one algorithm to be absolutely continuous with respect to another, a concept that we also introduce here.
Main definitions
-
AbsolutelyContinuous alg alg₀:algis absolutely continuous with respect toalg₀(also denotedalg ≪ₐ alg₀) when, in every situation, a set of actions with probability zero underalg₀also has probability zero underalg. Intuitively,algnever acts in a way thatalg₀would never act. -
density alg alg₀ n: a density function that allows obtaining the law of the history at timenunderalgfrom the law of the history at timenunderalg₀when they are interacting with the same environment andalg ≪ₐ alg₀.
Main results
-
absolutelyContinuous_map_hist: the law of the history at timenunderalgis absolutely continuous with respect to the law of the history at timenunderalg₀when they are interacting with the same environment andalg ≪ₐ alg₀. -
hasLaw_history_withDensity: the law of the history at timenunderalgis the law of the history at timenunderalg₀with densityalg.density alg₀ nwhen they are interacting with the same environment andalg ≪ₐ alg₀.
Module LeanMachineLearning.SequentialLearning.AlgorithmDensity contains 6 exposed declarations.
-
Learning.Algorithm.AbsolutelyContinuous -
Learning.Algorithm.«term_≪ₐ_» -
Learning.Algorithm.density -
Learning.Algorithm.measurable_density -
Learning.IsAlgEnvSeq.absolutelyContinuous_map_history -
Learning.IsAlgEnvSeq.hasLaw_history_withDensity