evolutionary-dynamics.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "http://www.offconvex.org/2016/03/07/evolution-markov-chains/\n",
    "http://www.pnas.org/content/111/29/10620.full.pdf\n",
    "http://theory.epfl.ch/vishnoi/Publications_files/VError.pdf\n",
    "http://www.math.harvard.edu/archive/153_fall_04/Additional_reading_material/evolutionary%20graph%20theory.pdf\n",
    "* Does evolutionary plasticity evolve?\n",
    "\n",
    "Population composed of different types!\n",
    "\n",
    "Evolution is just a random, local walk on a graph.\n",
    "$x_{t+1} = \\frac{AQx_t}{\\parallel AQx_t\\parallel_1}$\n",
    "where A is the fitness of the population (just diagonal)\n",
    "Q is the chance of mutating to some other x^i.\n",
    "\n",
    "I can image mutation propagating along a graph.\n",
    "\n",
    "So instead of a N-D continuous space we have a graph in the discrete case. Huh\n",
    "\n",
    "So mutation is a linear operator? Probably not? There should be interactions between the different species? What about epigenetics? Some genes are protected from mutation???\n",
    "\n",
    "How do we find A? How should the fitness be measured? Is it really a linear operator? Populations tend to grow exponentially?\n",
    "\n",
    "So is the number/proportion of different species.\n",
    "How can we represent the locality of similar species? Would need a tensor to represent every dimension of possible variability? Then we can flatten and apply AQ?\n",
    "This reminds me of quantum states?!?\n",
    "\n",
    "\n",
    "Mixing time. As we repeatedly apply a linear transform the eigenvalues tend to concentrate on ???.\n",
    "\n",
    "\n",
    "Asexual reproduction == convex optimisation?\n",
    "Sexual reproduction == non-convex optimisation!?\n",
    "What relation is there to exploration?\n",
    "\n",
    "\n",
    "* Relation between exaptation and transfer learning.\n",
    "* What about learning to learn/evolving to evolve better?\n",
    "* Which representations make sense!?\n",
    "* Novelty, fitness functions, unsupervised priors, ?\n",
    "* Evolution of memory/reasoning/cognition? (probably a separate topic)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}