SNN_two_layer.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# SNN with IF-Neurons\n",
    "\n",
    "This code is designed to demo a Spiking Neural Network propogating spikes between simple Integrate-and-Fire (IF) neurons.\n",
    "All parameters are integer only (to enable a simpler hardware implementation).\n",
    "Information is encoded with the Time-To-First-Spike (TTFS) scheme: information is encoded in the time that has passed between the arrival of a stimulus and the emission of the spike. A high activation is encoded as an earlier spike transmission and vice versa. Therefore, a high activation value is represented by the inverse TTFS.\n",
    "\n",
    "Note that the decoded values have a limeted accuracy due to the integer representation.\n",
    "\n",
    "There is no learning involved at this point."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### IF Neuron model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "class IFNeuron():\n",
    "    def __init__(self):        \n",
    "        \n",
    "        # neuron state variables\n",
    "        self.Vm        = []               # Neuron potential over time \n",
    "        self.spikes    = []               # Output spikes of the neuron\n",
    "        self.slope     = 0                # membrane potential slope\n",
    "        self.refrac    = 0                # timer to keep track of refrac time\n",
    "        self.idle      = True             # is neuron idle\n",
    "        self.encoding  = False            # is neuron encoding info\n",
    "        \n",
    "        # additional variables for plotting\n",
    "        self.time      = []               # Time steps (keeping track of time for plotting)\n",
    "        self.decoded   = []               # decoded inverse ttfs values\n",
    "        self.accuracy  = []               # accuracy of decoded value\n",
    "        self.enc_start = []               # start time of encoding window (for plotting)\n",
    "        self.input     = []               # keep track of weighted input spikes (for plotting)\n",
    "         \n",
    "        # IF Properties \n",
    "        self.t_refrac = 5                 # refractory period\n",
    "        self.thresh   = 255               # firing threshold\n",
    "        self.type     = 'Integrate and Fire'\n",
    "        \n",
    "        # initalize\n",
    "        self.Vm.append(0)\n",
    "        self.time.append(0)\n",
    "        self.spikes.append(0)\n",
    "    \n",
    "    def update(self, synapse_out):       # call to update Neuron\n",
    "        self.input.append(synapse_out)   # keep track of inputs (for plotting)\n",
    "        \n",
    "        Vm_new = 0                       # new membrane potential\n",
    "        slope_new = 0                    # new slope\n",
    "        neuron_output = 0                # output of neuron\n",
    "        \n",
    "        if not self.idle or synapse_out != 0: # calculate only when things are happening\n",
    "\n",
    "            if self.Vm[-1] >= self.thresh:   # exceed threshold\n",
    "                neuron_output = 1            # fire spike\n",
    "                self.refrac = self.t_refrac  # start refractory timer\n",
    "                self.encoding = False        # stop encoding window\n",
    "                self.decoded.append(self.thresh - (self.time[-1] - self.enc_start[-1]) + 1) # decode inverse TTFS\n",
    "\n",
    "                overshoot = self.Vm[-1] - self.thresh  # calc overshoot to estimate accuracy\n",
    "                self.accuracy.append((1 - overshoot / self.thresh) * 100) # accuracy is the amount of overshoot in relation to the threshold\n",
    "\n",
    "            else:                            # subthreshold, fire no spike\n",
    "                if synapse_out != 0 and self.refrac == 0 and not self.encoding: # spike occured, start encoding window!\n",
    "                    self.encoding = True                  # start encoding window\n",
    "                    self.enc_start.append(self.time[-1])  # remember start time of encoding window\n",
    "\n",
    "                if self.refrac == 0:                      # if not in refractory period,\n",
    "                    slope_new = synapse_out + self.slope  # update slope of membrane potential\n",
    "                    Vm_new = self.Vm[-1] + slope_new      # update membrane potential\n",
    "\n",
    "                else:                                     # if in refractory period, \n",
    "                    self.refrac = self.refrac - 1         # update refreactory timer\n",
    "        \n",
    "        \n",
    "        # save all params\n",
    "        self.spikes.append(neuron_output)\n",
    "        self.Vm.append(Vm_new)\n",
    "        self.slope = slope_new\n",
    "        self.idle = self.Vm[-1] == 0 and self.refrac == 0 and not self.encoding # check if neuron is idle\n",
    "        self.time.append(self.time[-1] + 1) # update time\n",
    "        "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Define neural network\n",
    "- Two neurons in layer 0\n",
    "- Two neurons in layer 1\n",
    "\n",
    "with the follwoing architecture:\n",
    "\n",
    "<img src=\"graph2.bmp\" width=\"300\" style=\"float: left;\"/>\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [],
   "source": [
    "# rerun this to re-initalize the neurons internal state\n",
    "# neurons on layer 0\n",
    "neuron0_l0 = IFNeuron()\n",
    "neuron1_l0 = IFNeuron()\n",
    "\n",
    "# neurons on layer 1\n",
    "neuron0_l1 = IFNeuron()\n",
    "neuron1_l1 = IFNeuron()\n",
    "\n",
    "# synaptic weights\n",
    "w0 = 1\n",
    "w1 = 2\n",
    "w00 = 5\n",
    "w01 = 7\n",
    "w10 = 12\n",
    "w11 = 4"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Define input spike trains"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [],
   "source": [
    "# input spike trains must have the same length\n",
    "spikes_0 = [0,0,0,1,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0] # input for neuron 0\n",
    "spikes_1 = [0,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1] # input for neuron 1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Run simulation\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "# helper function to check if network is idle\n",
    "def network_is_idle(list_of_neurons):\n",
    "    is_idle = True\n",
    "    for neuron in list_of_neurons:\n",
    "        is_idle = is_idle and neuron.idle\n",
    "    \n",
    "    return is_idle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "simulation took 190 time steps\n"
     ]
    }
   ],
   "source": [
    "time_step = 0\n",
    "\n",
    "# loop thru spike input trains and continue looping until the input is fully processed (i.e. all neurons are idle)\n",
    "while time_step < len(spikes_0) or not network_is_idle([neuron0_l0, neuron1_l0, neuron0_l1, neuron1_l1]):\n",
    "    \n",
    "    if time_step < len(spikes_0): # feed input spikes to neurons\n",
    "        # feed spike trains to the two neurons in layer 0\n",
    "        neuron0_l0.update(spikes_0[time_step] * w0)\n",
    "        neuron1_l0.update(spikes_1[time_step] * w1)\n",
    "    \n",
    "    else: # feed zero input to neurons \n",
    "        neuron0_l0.update(0)\n",
    "        neuron1_l0.update(0)\n",
    "\n",
    "    # calc weighted input of the two neurons in layer 1\n",
    "    input0_l1 = neuron0_l0.spikes[-1] * w00 + neuron1_l0.spikes[-1] * w10\n",
    "    input1_l1 = neuron0_l0.spikes[-1] * w01 + neuron1_l0.spikes[-1] * w11\n",
    "    \n",
    "    # feed the inputs into the neurons of layer 1\n",
    "    neuron0_l1.update(input0_l1)\n",
    "    neuron1_l1.update(input1_l1)\n",
    "    \n",
    "    time_step = time_step + 1\n",
    "\n",
    "print(f\"simulation took {time_step} time steps\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Plot the results of the simulation\n",
    "- Note how long encoding windows represent low activations (inverse TTFS)\n",
    "- Note how high activations have a lower accuracy than low activations. \n",
    "\n",
    "The accuracy can be increased by increasing the firing threshold at the cost of a longer simulation time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "decoded output activations: [126]\n",
      "decoded output activations: ['0.49'] (relative to the maximum possible activation)\n",
      "accuracy of decoded values: ['99.61'] %\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "decoded output activations: [183]\n",
      "decoded output activations: ['0.72'] (relative to the maximum possible activation)\n",
      "accuracy of decoded values: ['98.82'] %\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "decoded output activations: [234, 205]\n",
      "decoded output activations: ['0.92', '0.80'] (relative to the maximum possible activation)\n",
      "accuracy of decoded values: ['96.47', '100.00'] %\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "decoded output activations: [202]\n",
      "decoded output activations: ['0.79'] (relative to the maximum possible activation)\n",
      "accuracy of decoded values: ['98.82'] %\n"
     ]
    }
   ],
   "source": [
    "def plot_neuron(neuron_to_plot, name):\n",
    "    plt.rcParams['figure.figsize'] = [15, 2]\n",
    "    plt.ylabel('Membrane potential') \n",
    "    plt.xlabel('Time') \n",
    "    plt.title(f\"Membrane potential of {name} over time\")\n",
    "    \n",
    "    # plot membrane potential over time (blue)\n",
    "    plt.plot(neuron_to_plot.time, neuron_to_plot.Vm)\n",
    "    \n",
    "    # plot start of encoding window (green)\n",
    "    for elem in range(len(neuron_to_plot.enc_start)):\n",
    "        plt.axvline(x=neuron_to_plot.enc_start[elem], color='green', linestyle=(0, (5, 10)))\n",
    "    \n",
    "    # plot end of encoding window (red)\n",
    "    for elem in range(len(neuron_to_plot.spikes)):\n",
    "        if neuron_to_plot.spikes[elem] != 0:\n",
    "            plt.axvline(x=neuron_to_plot.time[elem], color='red', linestyle=(0, (5, 10))) \n",
    "    \n",
    "    # plot input spikes (black)\n",
    "    for elem in range(len(neuron_to_plot.input)):\n",
    "        if neuron_to_plot.input[elem] > 0:\n",
    "            plt.vlines(elem, ymin=0, ymax=neuron_to_plot.thresh/5, color='black', linewidth=1.2) # input spikes are all displayed with same size \n",
    "    \n",
    "    # plot threshold (grey)\n",
    "    plt.axhline(y=neuron_to_plot.thresh, color='grey', linestyle='dotted')\n",
    "    \n",
    "    plt.show()\n",
    "    \n",
    "    # print decoded activation\n",
    "    print(f\"decoded output activations: {neuron_to_plot.decoded}\")\n",
    "    #[a/b for a,b in zip(neuron_to_plot.decoded,neuron_to_plot.thresh)]\n",
    "    relative_activations = [a/neuron_to_plot.thresh for a in neuron_to_plot.decoded]\n",
    "    print(f\"decoded output activations: {['%.2f' % elem for elem in relative_activations]} (relative to the maximum possible activation)\")\n",
    "    print(f\"accuracy of decoded values: {['%.2f' % elem for elem in neuron_to_plot.accuracy]} %\") # display only 2 decimal places\n",
    "\n",
    "plot_neuron(neuron0_l0, \"neuron 0 layer 0\")\n",
    "plot_neuron(neuron1_l0, \"neuron 1 layer 0\")\n",
    "plot_neuron(neuron0_l1, \"neuron 0 layer 1\")\n",
    "plot_neuron(neuron1_l1, \"neuron 1 layer 1\")\n"
   ]
  }
 ],
 "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": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}