Skill chaining, a general method for skill acquisition in continuous reinforcement learning domains. Skill chaining incrementally builds a skill tree that allows an agent to reach a solution state from any of its start states by executing a sequence (or chain) of acquired skills. Skill chaining can adaptively break tasks that are too complex to be solved monolithically into sequences of subtasks that can be solved efficiently. The agent can then learn to sequence the resulting subtasks to solve the original problem. 

References[edit]

• Konidaris, George; Andrew Barto (2009). "Skill discovery in continuous reinforcement learning domains using skill chaining". Advances in Neural Information Processing Systems 22.