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

The binding problem is one of the fundamental challenges that prevent theartificial neural network (ANNs) from a compositional understanding of theworld like human perception, because disentangled and distributedrepresentations of generative factors can interfere and lead to ambiguity whencomplex data with multiple objects are presented. In this paper, we propose abrain-inspired hybrid neural network (HNN) that introduces temporal bindingtheory originated from neuroscience into ANNs by integrating spike timingdynamics (via spiking neural networks, SNNs) with reconstructive attention (byANNs). Spike timing provides an additional dimension for grouping, whilereconstructive feedback coordinates the spikes into temporal coherent states.Through iterative interaction of ANN and SNN, the model continuously bindsmultiple objects at alternative synchronous firing times in the SNN codingspace. The effectiveness of the model is evaluated on synthetic datasets ofbinary images. By visualization and analysis, we demonstrate that the bindingis explainable, soft, flexible, and hierarchical. Notably, the model is trainedon single object datasets without explicit supervision on grouping, butsuccessfully binds multiple objects on test datasets, showing its compositionalgeneralization capability. Further results show its binding ability in dynamicsituations.

Dance of SNN and ANN: Solving binding problem by combining spike timing and reconstructive attention