Parameter efficient dendritic-tree neurons outperform perceptrons

Ziwen Han, Evgeniya Gorobets, Pan Chen
7/2/2022
cs.NEcs.LG

No official code found. Best implementation: zw123han/DendriticNeuralNetwork (3 stars, MIT, last commit 2022-07-10).

Abstract

Biological neurons are more powerful than artificial perceptrons, in part due to complex dendritic input computations. Inspired to empower the perceptron with biologically inspired features, we explore the effect of adding and tuning input branching factors along with input dropout. This allows for parameter efficient non-linear input architectures to be discovered and benchmarked. Furthermore, we present a PyTorch module to replace multi-layer perceptron layers in existing architectures. Our initial experiments on MNIST classification demonstrate the accuracy and generalization improvement of dendritic neurons compared to existing perceptron architectures.

View on arXivView PDF

Code Implementations(1)

zw123han/DendriticNeuralNetworkCommunity100%
MIT

Repo for dendritic neural network experiments

30HTML, PythonApr 13, 20224 years agoMIT

Similar Papers

Dendrites endow artificial neural networks with accurate, robust and parameter-efficient learning

86% match2024

Dendritic Integration Based Quadratic Neural Networks Outperform Traditional Aritificial Ones

86% match2023

Neuromorphic Online Clustering and Its Application to Spike Sorting

85% match2025

Leveraging dendritic properties to advance machine learning and neuro-inspired computing

85% match2023

NSPDI-SNN: An efficient lightweight SNN based on nonlinear synaptic pruning and dendritic integration

84% match2025

Perforated Backpropagation: A Neuroscience Inspired Extension to Artificial Neural Networks

83% match2025

Cite this paper

@article{han2022parameter,
  title  = {Parameter efficient dendritic-tree neurons outperform perceptrons},
  author = {Ziwen Han and Evgeniya Gorobets and Pan Chen},
  year   = {2022},
  eprint = {2207.00708},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2207.00708v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2207.00708.svg)](https://codeofpaper.com/paper/2207.00708)

Discussion