Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Protein pathways as a catalyst to directed evolution of the topology of artificial neural networks

Oscar Lao, Konstantinos Zacharopoulos, Apostolos Fournaris, Rossano Schifanella, Ioannis Arapakis

TL;DR

A paradigm shift on evolving Artificial Neural Networks (ANNs) is proposed towards a new bio-inspired design that is grounded on the structural properties, interactions, and dynamics of protein networks (PNs): the Artificial Protein Network (APN).

Abstract

In the present article, we propose a paradigm shift on evolving Artificial Neural Networks (ANNs) towards a new bio-inspired design that is grounded on the structural properties, interactions, and dynamics of protein networks (PNs): the Artificial Protein Network (APN). This introduces several advantages previously unrealized by state-of-the-art approaches in NE: (1) We can draw inspiration from how nature, thanks to millions of years of evolution, efficiently encodes protein interactions in the DNA to translate our APN to silicon DNA. This helps bridge the gap between syntax and semantics observed in current NE approaches. (2) We can learn from how nature builds networks in our genes, allowing us to design new and smarter networks through EA evolution. (3) We can perform EA crossover/mutation operations and evolution steps, replicating the operations observed in nature directly on the genotype of networks, thus exploring and exploiting the phenotypic space, such that we avoid getting trapped in sub-optimal solutions. (4) Our novel definition of APN opens new ways to leverage our knowledge about different living things and processes from biology. (5) Using biologically inspired encodings, we can model more complex demographic and ecological relationships (e.g., virus-host or predator-prey interactions), allowing us to optimise for multiple, often conflicting objectives.

Protein pathways as a catalyst to directed evolution of the topology of artificial neural networks

TL;DR

A paradigm shift on evolving Artificial Neural Networks (ANNs) is proposed towards a new bio-inspired design that is grounded on the structural properties, interactions, and dynamics of protein networks (PNs): the Artificial Protein Network (APN).

Abstract

In the present article, we propose a paradigm shift on evolving Artificial Neural Networks (ANNs) towards a new bio-inspired design that is grounded on the structural properties, interactions, and dynamics of protein networks (PNs): the Artificial Protein Network (APN). This introduces several advantages previously unrealized by state-of-the-art approaches in NE: (1) We can draw inspiration from how nature, thanks to millions of years of evolution, efficiently encodes protein interactions in the DNA to translate our APN to silicon DNA. This helps bridge the gap between syntax and semantics observed in current NE approaches. (2) We can learn from how nature builds networks in our genes, allowing us to design new and smarter networks through EA evolution. (3) We can perform EA crossover/mutation operations and evolution steps, replicating the operations observed in nature directly on the genotype of networks, thus exploring and exploiting the phenotypic space, such that we avoid getting trapped in sub-optimal solutions. (4) Our novel definition of APN opens new ways to leverage our knowledge about different living things and processes from biology. (5) Using biologically inspired encodings, we can model more complex demographic and ecological relationships (e.g., virus-host or predator-prey interactions), allowing us to optimise for multiple, often conflicting objectives.
Paper Structure (5 sections, 6 figures)

This paper contains 5 sections, 6 figures.

Table of Contents

  1. Introduction
  2. Evolution as a powerful meta-optimisation algorithm

Figures (6)

  • Figure 1: Evolution serves as a metaheuristic for optimization. Genetic diversity within a population leads to varied phenotypes influenced by environmental factors. Through reproduction, genetic material merges, creating novel phenotypes. This process involves navigating phenotypic space for fitness assessment and systematically exploring and adapting to the environment. A genetic algorithm mimics this process.
  • Figure 2: Example of an encoding-decoding pathway of Artificial Protein Network (APN) to silicon chromosome
  • Figure 3: Biological similarities of PNs and NNs. The genetic blueprint of a species contains both NNs and PNs. Neurons are at the core of NNs, forming neural tissue up to the brain. Decoding NNs from DNA is difficult as NNs are shaped by environment and development. Directed signed PNs use proteins as basic units, with functions directly dictated by DNA, that enable PPIs and forming PNs that contribute to cell functionality.
  • Figure 4: Left side: (a) A DL model for emulating the biological non-linear decoding of the genetic information present in the DNA to generate a network. Right side: (b) Genotype to phenotype mapping and (c) structural mutilation examples, as implemented in NEAT Kenneth2002
  • Figure 5: An artificial neural network is mirrored by particular connections between proteins.
  • ...and 1 more figures