17546022. INTEGRATED AI PLANNERS AND RL AGENTS THROUGH AI PLANNING ANNOTATION IN RL simplified abstract (INTERNATIONAL BUSINESS MACHINES CORPORATION)

INTEGRATED AI PLANNERS AND RL AGENTS THROUGH AI PLANNING ANNOTATION IN RL

Organization Name

INTERNATIONAL BUSINESS MACHINES CORPORATION

Inventor(s)

Junkyu Lee of San Diego CA (US)

Michael Katz of Goldens Bridge NY (US)

Shirin Sohrabi Araghi of Briarcliff manor NY (US)

Don Joven Ravoy Agravante of Tokyo (JP)

Miao Liu of Ossining NY (US)

Tamir Klinger of Brooklyn NY (US)

Murray Scott Campbell of Yorktown Heights NY (US)

INTEGRATED AI PLANNERS AND RL AGENTS THROUGH AI PLANNING ANNOTATION IN RL - A simplified explanation of the abstract

This abstract first appeared for US patent application 17546022 titled 'INTEGRATED AI PLANNERS AND RL AGENTS THROUGH AI PLANNING ANNOTATION IN RL

Simplified Explanation

The abstract describes a computer-implemented method that integrates an AI planner and a reinforcement learning agent. This integration is achieved through AI planning annotation in RL (PaRL). Here are the key points:

• The method starts by identifying an RL problem.
• A description of a Markov decision process (MDP) in an RL environment is received.
• An RL task is generated based on the MDP description to solve the RL problem.
• An AI planning model described in a planning language is received.
• The method maps the state spaces from the MDP states to the AI planning states of the AI planning model.
• The RL task is generated with an AI planning task from the mapping, resulting in a PaRL task.

Potential Applications

• This technology can be applied in various fields where AI planning and reinforcement learning are used together, such as robotics, autonomous systems, and game playing.
• It can be used to enhance decision-making processes in complex environments by combining the strengths of AI planning and reinforcement learning.

Problems Solved

• Integrating AI planning and reinforcement learning can be challenging due to the differences in their approaches and representations.
• This method solves the problem of effectively combining AI planning and reinforcement learning by using AI planning annotation in RL.

Benefits

• By integrating AI planning and reinforcement learning, this method can improve the performance and efficiency of decision-making systems.
• It allows for better utilization of the strengths of both AI planning and reinforcement learning techniques.
• The method provides a systematic approach to integrate AI planning and reinforcement learning, making it easier to develop and deploy intelligent systems.

Original Abstract Submitted

A computer-implemented method of integrating an Artificial Intelligence (AI) planner and a reinforcement learning (RL) agent through AI planning annotation in RL (PaRL) includes identifying an RL problem. A description received of a Markov decision process (MDP) having a plurality of states in an RL environment is used to generate an RL task to solve the RL problem. An AI planning model described in a planning language is received, and mapping state spaces from the MDP states in the RL environment to AI planning states of the AI planning model is performed. The RL task is generated with an AI planning task from the mapping to generate a PaRL task.