Mar 5, 2020

Autoencoder networks extract latent variables and encode these variables in their connectomes

BioRxiv : the Preprint Server for Biology
Matthew FarrellEric Shea-Brown

Abstract

Spectacular advances in imaging and data processing techniques are revealing a wealth of information about brain connectomes. This raises an exciting scientific opportunity: to infer the underlying circuit function from the structure of its connectivity. A potential roadblock, however, is that -- even with well constrained neural dynamics -- there are in principle many different connectomes that could support a given computation. Here, we define a tractable setting in which the problem of inferring circuit function from circuit connectivity can be analyzed in detail: the function of input compression and reconstruction, in an autoencoder network with a single hidden layer. Here, in general there is substantial ambiguity in the weights that can produce the same circuit function, because largely arbitrary changes to ''input'' weights can be undone by applying the inverse modifications to the ''output'' weights. However, we use mathematical arguments and simulations to show that adding simple, biologically motivated regularization of connectivity resolves this ambiguity in an interesting way: weights are constrained such that the latent variable structure underlying the inputs can be extracted from the weights by using nonlinear d...Continue Reading

  • References
  • Citations

References

  • We're still populating references for this paper, please check back later.
  • References
  • Citations

Citations

  • This paper may not have been cited yet.

Mentioned in this Paper

Gene Circuits
Diagnostic Imaging
Problem
Motivation
Neural Network Simulation
Structure
Analysis
Simulation
Reduction - Action
Reconstructive Surgical Procedures

Related Feeds

BioRxiv & MedRxiv Preprints

BioRxiv and MedRxiv are the preprint servers for biology and health sciences respectively, operated by Cold Spring Harbor Laboratory. Here are the latest preprint articles (which are not peer-reviewed) from BioRxiv and MedRxiv.

Allen Institute for Brain Science Network

The Allen Institute for Brain Science Network is a not-for-profit biomedical research organization that provides open access to multiple neuroscience tools and resources. Find the latest research from tjr Allen Institute for Brain Science Network here.