EW‐CACTUs‐MAML: A Robust Metalearning System for Rapid Classification on a Large Number of Tasks

Abstract

This study aims to develop a robust metalearning system for rapid classification on a large number of tasks. The model-agnostic metalearning (MAML) with the CACTUs method (clustering to automatically construct tasks for unsupervised metalearning) is improved as EW-CACTUs-MAML after integrated with the entropy weight (EW) method. Few-shot mechanisms are introduced in the deep network for efficient learning of a large number of tasks. The process of implementation is theoretically interpreted as “gene intelligence.” Validation of EW-CACTUs-MAML on a typical dataset (Omniglot) indicates an accuracy of 97.42%, performing better than CACTUs-MAML (validation accuracy = 97.22%). At the end of this paper, the availability of our thoughts to improve another metalearning system (EW-CACTUs-ProtoNets) is also preliminarily discussed based on a cross-validation on another typical dataset (Miniimagenet).

A new method of feature splicing based on wavelet transform for recognition of HRRP with noise

“…Taking the missing data as the test set, corresponding to the situation that some attitude angle data are missing. Under the condition of small samples with missing attitude angle data, MAML algorithm [13] is used for training and recognition. In the training stage, α angular domains are extracted from each kind of target, and k + k samples from each angular domain constitute a task, and a certain number of tasks form a batch to update the parameter θ.…”