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 Plus annual

Presents the access of premium content as premium feature

Premium Content

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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Combinatorial generalization - the ability to understand and produce novelcombinations of already familiar elements - is considered to be a core capacityof the human mind and a major challenge to neural network models. A significantbody of research suggests that conventional neural networks can't solve thisproblem unless they are endowed with mechanisms specifically engineered for thepurpose of representing symbols. In this paper we introduce a novel way ofrepresenting symbolic structures in connectionist terms - the vectors approachto representing symbols (VARS), which allows training standard neuralarchitectures to encode symbolic knowledge explicitly at their output layers.In two simulations, we show that neural networks not only can learn to produceVARS representations, but in doing so they achieve combinatorial generalizationin their symbolic and non-symbolic output. This adds to other recent work thathas shown improved combinatorial generalization under specific trainingconditions, and raises the question of whether specific mechanisms or trainingroutines are needed to support symbolic processing.

Training neural networks to encode symbols enables combinatorial generalization