Searching for algorithmic principles of adaptive network design and load balancing in the nervous system

Abstract

Biological systems must constantly solve what we think of as computational problems: foraging slime molds design networks to optimize food transport, and cells use delta-notch signaling to perform leader election when determining cell fate. Understanding biological processes in this way requires mapping the process to an algorithm that solves an optimization problem under some constraints. The goal of this proposal is to establish links between fundamental neurobiological processes involved in learning and adaptation in the brain, and algorithms designed to compute in uncertain, changing environments. This will lead to advances in both computer science, by developing new bio-inspired algorithms for a variety of communication- and resource-constrained networking problems, and in neuroscience, by experimentally determining local computational rules implemented by individual neurons and synapses and relating them to systems-level information processing. We will focus on two problems: (1) synaptic pruning and network design, and (2) neural homeostasis and load balancing.

Document Details

Document Type
DoD Grant Award
Publication Date
Oct 15, 2018
Source ID
W911NF1710045

Entities

People

  • Saket Navlakha

Organizations

  • Army Contracting Command
  • Salk Institute for Biological Studies
  • United States Army

Tags

Readers

  • Neural Network Machine Learning.
  • Operations Research
  • Systems Analysis and Design