US 12,333,434 B2
Meta imitation learning with structured skill discovery
Wenchao Yu, Plainsboro, NJ (US); Wei Cheng, Princeton Junction, NJ (US); Haifeng Chen, West Windsor, NJ (US); and Yiwei Sun, State College, PA (US)
Assigned to NEC Corporation, Tokyo (JP)
Filed by NEC Laboratories America, Inc., Princeton, NJ (US)
Filed on Oct. 11, 2023, as Appl. No. 18/484,816.
Application 18/484,816 is a continuation of application No. 17/391,427, filed on Aug. 2, 2021.
Claims priority of provisional application 63/084,035, filed on Sep. 28, 2020.
Claims priority of provisional application 63/067,009, filed on Aug. 18, 2020.
Prior Publication US 2024/0046092 A1, Feb. 8, 2024
This patent is subject to a terminal disclaimer.
Int. Cl. G06N 3/08 (2023.01); G06N 20/00 (2019.01)
CPC G06N 3/08 (2013.01) [G06N 20/00 (2019.01)] 7 Claims
OG exemplary drawing
 
1. A method for acquiring skills through imitation learning, the method comprising:
learning behaviors or tasks, by an agent, from state-action pairs of medical treatment for a given disease by:
learning to decompose the state-action pairs into segments, via a segmentation component, the segments corresponding to skills that are transferrable across different tasks;
learning relationships between the skills;
employing, via a graph generator, a graph neural network for learning implicit structures of the skills from the state-action pairs 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 by optimizing an objective function:

OG Complex Work Unit Math
wherein pπθ is a generated policy with parameters πθ, pπE is an expert policy, custom character is a distance function, H is a Shannon entropy function, c is a set of skills, X is an implicit structure, g is a segmentation, and s is a state; and
providing a medication to a patient in accordance with the generated policies to treat the given disease.