KIP publications

 
year 2020
author(s) Akos F. Kungl, Dominik Dold, Oskar Riedler, Mihai A. Petrovici, Walter Senn
title Deep reinforcement learning for time-continuous substrates
KIP-Nummer HD-KIP 20-16
KIP-Gruppe(n) F9
document type Paper
source Neuro-Inspired Computational Elements Workshop (NICE), 2020 Heidelberg, Germany
Abstract (en)

To achieve their goal of realizing fast and energy-efficient learning, neuromorphic systems require computationally powerful models that obey the constraints imposed by a physical implementation of neural network structure and dynamics, such as the inevitability of relaxation times or the locality of plasticity. In this work, we provide a first-principles derivation of a mechanistic model for cortical computation based on the premise of "neuronal least action".  The resulting time-continuous neuron and synapse dynamics realize gradient-descent learning through error backpropagation both in supervised and in reinforcement learning scenarios. In particular, the derived equations of motion reproduce well-established microscopic phenomena such as neuronal leaky integration of afferent signals, while enabling synaptic learning using only locally available information. Our principled framework can thus serve as a starting point for hardware-focused models of highly efficient time-continuous learning.

bibtex
@conference{kungl2020deep,
  author   = {Kungl, Akos F. and Dold, Dominik and Riedler, Oskar and Petrovici, Mihai A. and Senn, Walter},
  title    = {Deep reinforcement learning for time-continuous substrates},
  booktitle = {},
  year     = {2020},
  volume   = {},
  pages    = {}
}
Datei pdf
up
KIP - Bibliothek
Im Neuenheimer Feld 227
Raum 3.402
69120 Heidelberg