Title:
METHOD OF NEURAL ARCHITECTURE SEARCH USING CONTINUOUS ACTION REINFORCEMENT LEARNING
Document Type and Number:
WIPO Patent Application WO/2022/068934
Kind Code:
A1
Abstract:
A method and system for generating neural architectures to perform a particular task. An actor neural network, as part of a continuous action reinforcement learning (RL) agent, generates a randomized continuous actions parameters to encourage exploration of a search space to generate candidate architectures without bias. The continuous action parameters are discretized and applied to a search space to generate candidate architectures, the performance of which for performing the particular task is evaluated. Corresponding reward and state are determined based on the performance. A critic neural network, as part of the continuous action RL agent, learns a mapping of the continuous action to a reward using modified Deep Deterministic Policy Gradient (DDPG) with quantile loss function by sampling a list of top performing architectures. The actor neural network is updated with the learned mapping.
More Like This:
Inventors:
SALAMEH MOHAMMAD (CA)
MILLS KEITH GEORGE (CA)
NIU DI (CA)
MILLS KEITH GEORGE (CA)
NIU DI (CA)
Application Number:
PCT/CN2021/122384
Publication Date:
April 07, 2022
Filing Date:
September 30, 2021
Export Citation:
Assignee:
HUAWEI TECH CO LTD (CN)
International Classes:
G06N3/08
Domestic Patent References:
WO2020160252A1 | 2020-08-06 | |||
WO2019007388A1 | 2019-01-10 |
Foreign References:
CN110651279A | 2020-01-03 | |||
US20200082275A1 | 2020-03-12 | |||
US20200293883A1 | 2020-09-17 |
Download PDF:
Previous Patent: COMPOUNDS AND METHODS OF TREATING DISEASES
Next Patent: ELEVATOR RESOURCE SCHEDULING METHOD AND APPARATUS
Next Patent: ELEVATOR RESOURCE SCHEDULING METHOD AND APPARATUS