Intelligent control with hierarchical stacked neural networks

This invention describes a method for an intelligent agent to learn how to control its actions within an environment. It works by first estimating how the environment changes and modeling it as a step-by-step decision process. The agent then uses a multi-layered artificial neural network, where different layers are trained for different purposes, to decide its actions, continuously learning and improving its strategy by observing the environment's response and receiving feedback. The claims specifically detail processing a 'noise vector' within these hierarchical layers to refine the policy.

Why it matters: The core concepts of hierarchical neural networks and reinforcement learning for intelligent control, particularly with Markov Decision Process modeling, have seen significant advancements in deep learning research and computational power since 2018. This progress makes the training and deployment of such complex, adaptive control systems, including the handling of 'noise vectors' in hierarchical structures, far more feasible and effective today.

Status: Not in forceListed as no longer active. The specific reason is not in the record we hold.
How hard to build: SpecializedSoftware logic, neural network training

3 AI reinvention directions for this patent

See all three

AI gives you a few directions you could take this. Pick one, and we check whether your version is different enough to patent, then write the filing.

Reinvent this with AI

Where this one came from

Every morning we go through millions of abandoned patents and pick out 15 worth your time. This was one of today’s.

This one

+10