Emergent Language Symbolic Autoencoder (ELSA) with Weak Supervision to Model Hierarchical Brain Networks

arXiv:2404.10031v1 Announce Type: cross
Abstract: Brain networks display a hierarchical organization, a complexity that poses a challenge for existing deep learning models, often structured as flat classifiers, leading to difficulties in interpretability and the 'black box' issue. To bridge this gap, we propose a novel architecture: a symbolic autoencoder informed by weak supervision and an Emergent Language (EL) framework. This model moves beyond traditional flat classifiers by producing hierarchical clusters and corresponding imagery, subsequently represented through symbolic sentences to improve …

abstract architecture arxiv autoencoder black box box brain bridge challenge classifiers complexity cs.ai cs.lg deep learning flat gap hierarchical interpretability issue language networks novel organization q-bio.nc supervision type