An area of machine learning concerned with creation and modification of software agent’s knowledge about effects and preconditions of the actions that can be executed within its environment. This knowledge is usually represented in logic-based action description language and used as the input for automated planners.
