20240046091. META IMITATION LEARNING WITH STRUCTURED SKILL DISCOVERY simplified abstract (NEC Laboratories America, Inc.)

From WikiPatents
Jump to navigation Jump to search

META IMITATION LEARNING WITH STRUCTURED SKILL DISCOVERY

Organization Name

NEC Laboratories America, Inc.

Inventor(s)

Wenchao Yu of Plainsboro NJ (US)

Wei Cheng of Princeton Junction NJ (US)

Haifeng Chen of West Windsor NJ (US)

Yiwei Sun of State College PA (US)

META IMITATION LEARNING WITH STRUCTURED SKILL DISCOVERY - A simplified explanation of the abstract

This abstract first appeared for US patent application 20240046091 titled 'META IMITATION LEARNING WITH STRUCTURED SKILL DISCOVERY

Simplified Explanation

The patent application describes a method for acquiring skills through imitation learning using a meta imitation learning framework with structured skill discovery (MILD). The method involves an agent learning behaviors or tasks from demonstrations. The agent learns to decompose the demonstrations into segments using a segmentation component. These segments correspond to skills that can be transferred across different tasks. The agent also learns relationships between these transferable skills using a graph neural network generated by a graph generator. The neural network learns the implicit structures of the skills from the demonstrations to define structured skills. Finally, the agent generates policies from the structured skills to acquire them for application to one or more target tasks.

  • The method involves an agent learning behaviors or tasks from demonstrations.
  • The demonstrations are decomposed into segments using a segmentation component.
  • The segments correspond to skills that can be transferred across different tasks.
  • Relationships between the transferable skills are learned using a graph neural network.
  • The neural network defines structured skills by learning the implicit structures from the demonstrations.
  • Policies are generated from the structured skills to acquire them for application to target tasks.

Potential applications of this technology:

  • Robotic automation: The method can be applied to teach robots new skills by imitating human demonstrations, allowing them to perform complex tasks autonomously.
  • Virtual assistants: Virtual assistants can learn new skills by imitating human demonstrations, enabling them to provide more advanced and personalized assistance.
  • Gaming: The method can be used to train game characters to learn new skills by imitating expert players, enhancing the realism and adaptability of in-game characters.

Problems solved by this technology:

  • Skill acquisition: The method provides a systematic approach for agents to acquire new skills by imitating demonstrations, enabling them to perform a wide range of tasks.
  • Transfer learning: By identifying transferable skills and learning their relationships, the method allows agents to apply acquired skills to different tasks, reducing the need for extensive retraining.
  • Implicit skill structure discovery: The graph neural network learns the implicit structures of skills from demonstrations, providing a structured representation that facilitates skill acquisition and policy generation.

Benefits of this technology:

  • Efficiency: The method allows agents to learn new skills from demonstrations, reducing the time and effort required for manual programming or trial-and-error learning.
  • Adaptability: By learning transferable skills and their relationships, agents can apply acquired skills to different tasks, making them more versatile and adaptable.
  • Generalization: The structured representation of skills enables agents to generalize their knowledge and apply it to new situations, improving their performance and problem-solving abilities.


Original Abstract Submitted

a method for acquiring skills through imitation learning by employing a meta imitation learning framework with structured skill discovery (mild) is presented. the method includes learning behaviors or tasks, by an agent, from demonstrations: by learning to decompose the demonstrations into segments, via a segmentation component, the segments corresponding to skills that are transferrable across different tasks, learning relationships between the skills that are transferrable across the different tasks, employing, via a graph generator, a graph neural network for learning implicit structures of the skills from the demonstrations to define structured skills, and generating policies from the structured skills to allow the agent to acquire the structured skills for application to one or more target tasks.

Retrieved from "http://wikipatents.org/index.php?title=20240046091._META_IMITATION_LEARNING_WITH_STRUCTURED_SKILL_DISCOVERY_simplified_abstract_(NEC_Laboratories_America,_Inc.)&oldid=32719"