English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

The performance of gradient-based optimization strategies depends heavily onthe initial weights of the parametric model. Recent works show that there existweight initializations from which optimization procedures can find thetask-specific parameters faster than from uniformly random initializations andthat such a weight initialization can be learned by optimizing a specific modelarchitecture across similar tasks via MAML (Model-Agnostic Meta-Learning).Current methods are limited to populations of classification tasks that sharethe same number of classes due to the static model architectures used duringmeta-learning. In this paper, we present HIDRA, a meta-learning approach thatenables training and evaluating across tasks with any number of targetvariables. We show that Model-Agnostic Meta-Learning trains a distribution forall the neurons in the output layer and a specific weight initialization forthe ones in the hidden layers. HIDRA explores this by learning one masterneuron, which is used to initialize any number of output neurons for a newtask. Extensive experiments on the Miniimagenet and Omniglot data setsdemonstrate that HIDRA improves over standard approaches while generalizing totasks with any number of target variables. Moreover, our approach is shown torobustify low-capacity models in learning across complex tasks with a highnumber of classes for which regular MAML fails to learn any feasibleinitialization.

HIDRA: Head Initialization across Dynamic targets for Robust Architectures