{ "cells": [ { "metadata": {}, "cell_type": "markdown", "source": "# Reservoir computing for time series prediction of Mackey-Glass sequence" }, { "cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [], "source": [ "import svetlanna as sv\n", "from svetlanna.units import ureg\n", "import torch\n", "import tqdm\n", "from svetlanna.detector import Detector\n", "from svetlanna.simulation_parameters import SimulationParameters\n", "from collections import deque\n", "from typing import Sequence, TypeVar\n", "import matplotlib.pyplot as plt" ] }, { "cell_type": "markdown", "metadata": {}, "source": "## Mackey-Glass" }, { "cell_type": "markdown", "metadata": {}, "source": [ "The Mackey-Glass equation (DOI: 10.1126/science.267326, 4b) is given by:\n", "$$\n", "\\frac{dx}{dt} = \\frac{\\beta \\theta^n x(t-\\tau)}{\\theta^n + x^n(t-\\tau)} - \\gamma x(t)\n", "$$\n", "In discrete form:\n", "$$\n", "x[t+1] = \\frac{\\beta \\theta^n x[t-\\tau]}{\\theta^n + x^n[t-\\tau]} - \\gamma x[t]\n", "$$" ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [], "source": [ "T = TypeVar('T', bound=float)\n", "\n", "def mackey_glass_generator(\n", " x0: Sequence[T],\n", " n: float,\n", " beta: float,\n", " gamma: float,\n", " theta: float\n", "):\n", " queue: deque[T] = deque(x0)\n", "\n", " while True:\n", " x = queue.popleft()\n", " x_prev = queue[-1]\n", " new_x = beta * theta**n * x / (theta**n + x**n) - gamma * x_prev\n", "\n", " queue.append(new_x)\n", " yield new_x" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Reservoir\n", "\n", "A simple reservoir model (the main idea is explained in https://doi.org/10.1364/OE.20.022783):\n", "$$\n", " x_{out}[i] = F_{NL}(\\beta x_{in}[i] + \\alpha F_{D}(x_{out}[i-\\tau]))\n", "$$\n", "where $x_{in}$ is the input signal, $\\tau$ is the time delay, $F_{NL}$ is the nonlinear element, $F_{D}$ is the delay element, $\\alpha$ is the input gain, and $\\beta$ is the feedback gain.\n", "\n", "In this example, $F_{NL}$ is a diffractive layer with a random mask, and $F_{D}$ consists of nonlinear element with a nonlinear amplitude response\n", "$$\n", "|\\vec{E}_{out}|=\\dfrac{|\\vec{E}_{in}|^2}{1 + |\\vec{E}_{in}|^2}\n", "$$\n", "and free space." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "L = 1 * ureg.cm # size of the simulation domain\n", "N = 5 # mesh size\n", "\n", "\n", "sim_params = sv.SimulationParameters(\n", " {\n", " 'W': torch.linspace(-L/2, L/2, N),\n", " 'H': torch.linspace(-L/2, L/2, N),\n", " 'wavelength': 1 * ureg.um\n", " }\n", ")\n", "\n", "# Diffractive layer\n", "difflayer = sv.elements.DiffractiveLayer(\n", " sim_params,\n", " mask=torch.rand((N, N)),\n", " mask_norm=1\n", ")\n", "\n", "# Element with nonlinear response\n", "nonlinear_element = sv.elements.NonlinearElement(\n", " simulation_parameters=sim_params,\n", " response_function=lambda x: x**2 / (1 + x**2)\n", ")\n", "\n", "# Free space\n", "fs = sv.elements.FreeSpace(\n", " sim_params,\n", " distance=2 * ureg.cm,\n", " method='AS'\n", ")\n", "\n", "# The reservoir system\n", "reservoir = sv.elements.SimpleReservoir(\n", " sim_params,\n", " nonlinear_element=difflayer,\n", " delay_element=sv.LinearOpticalSetup(\n", " [nonlinear_element, fs]\n", " ),\n", " feedback_gain=0.8,\n", " input_gain=1.,\n", " delay=10\n", ")\n", "\n", "# Detector\n", "detector = Detector(\n", " sim_params\n", ")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "To enable masking of the input signal, one should create a custom module that applies masking in the forward method and then calculates the average of the detector output over a single symbol." ] }, { "cell_type": "code", "execution_count": 27, "metadata": {}, "outputs": [], "source": [ "class ReservoirSystem(torch.nn.Module):\n", " def __init__(\n", " self,\n", " simulation_parameters: SimulationParameters,\n", " reservoir: sv.elements.SimpleReservoir,\n", " detector: sv.elements.Element,\n", " mask: torch.Tensor\n", " ):\n", " super().__init__()\n", "\n", " assert len(mask.shape) == 1\n", "\n", " self.mask = mask\n", " self.simulation_parameters = simulation_parameters\n", " self.reservoir = reservoir\n", " self.detector = detector\n", "\n", " def forward(self, x) -> torch.Tensor:\n", "\n", " # before new sequence one should drop the wavefronts\n", " # stored in the feedback line\n", " self.reservoir.drop_feedback_queue()\n", "\n", " res = torch.empty(\n", " (\n", " x.shape[0],\n", " self.simulation_parameters.axes_size('W')[0] * self.simulation_parameters.axes_size('H')[0]\n", " )\n", " )\n", " mask_size = self.mask.shape[0]\n", "\n", " # For each symbol in the sequence\n", " for i, u in enumerate(x):\n", " masked_input = u * self.mask\n", "\n", " detector_signal = 0\n", " # For each mask frame in the mask\n", " for u_masked in masked_input:\n", " # Encode the signal in the amplitude of incident field\n", " wf = u_masked * sv.Wavefront.gaussian_beam(\n", " self.simulation_parameters,\n", " waist_radius=0.2 * ureg.cm\n", " )\n", " # Calculate the propagation through the reservoir\n", " wf = self.reservoir(wf)\n", " detector_signal += self.detector(wf)\n", "\n", " # Calculate average signal on the detector during the symbol\n", " # The output of this module is a vector of real value\n", " res[i] = detector_signal.ravel() / mask_size\n", "\n", " return res" ] }, { "cell_type": "markdown", "metadata": {}, "source": "## Dataset" }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "MG_TRAIN_SIZE = 200\n", "\n", "MG_X_PREHEAT_NUM = 100 # Number of preheat symbols\n", "MG_X_NUM = MG_X_PREHEAT_NUM + 200\n", "MG_DATASET_PREHEAT_NUM = 200 # preheat for dataset" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "First, we create a dataset from the Mackey-Glass sequence." ] }, { "cell_type": "code", "execution_count": 42, "metadata": {}, "outputs": [], "source": [ "class MGDataset(torch.utils.data.Dataset):\n", " def __init__(\n", " self,\n", " size: int,\n", " initial_step: int\n", " ):\n", " u = []\n", " u_generator = mackey_glass_generator(\n", " [1, 2, 1., 1, 1, 3, 3, 1, 2, 1, 0, 1, 1, 1, 1, 1],\n", " n=12,\n", " beta=0.3,\n", " gamma=-0.9,\n", " theta=4\n", " )\n", "\n", " for _ in range(initial_step):\n", " next(u_generator)\n", "\n", " for _ in range(MG_X_NUM + size + 1):\n", " u.append(next(u_generator))\n", "\n", " self.u = u\n", " self._size = size\n", "\n", " def __len__(self):\n", " return self._size\n", "\n", " def __getitem__(self, idx):\n", " x = torch.tensor(self.u[idx : idx + MG_X_NUM])\n", " y = torch.tensor(self.u[idx + 1 : idx + MG_X_NUM + 1])\n", " return x, y" ] }, { "cell_type": "code", "execution_count": 43, "metadata": {}, "outputs": [], "source": [ "mg_dataset = MGDataset(\n", " size=MG_TRAIN_SIZE,\n", " initial_step=MG_DATASET_PREHEAT_NUM\n", ")" ] }, { "cell_type": "code", "execution_count": 44, "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "plt.plot(mg_dataset.u, '-o', markersize=3)\n", "plt.xlabel('symbol index')\n", "plt.ylabel('$x_{in}$')\n", "plt.xlim(0, 200)\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Then, we process this dataset through the reservoir system.\n", "The final dataset will contain the input signal $x_{in}$, the signal vector at the detector after propagation through the reservoir system $y_{reservoir}$ and the required prediction of the next signal $y$:" ] }, { "cell_type": "code", "execution_count": 45, "metadata": {}, "outputs": [], "source": [ "class ReservoirResponseDataset(torch.utils.data.Dataset):\n", " def __init__(\n", " self,\n", " dataset: MGDataset,\n", " reservoir_system: ReservoirSystem\n", " ):\n", " self.xy = []\n", " \n", " for i in tqdm.tqdm(range(len(dataset))):\n", " x, y = dataset[i]\n", " y_reservoir = reservoir_system(x)\n", " self.xy.append((x, y_reservoir, y))\n", "\n", " def __len__(self):\n", " return len(self.xy)\n", "\n", " def __getitem__(self, idx):\n", " return self.xy[idx]" ] }, { "cell_type": "code", "execution_count": 46, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 200/200 [00:25<00:00, 7.87it/s]\n" ] } ], "source": [ "reservoir_system = ReservoirSystem(\n", " sim_params,\n", " reservoir,\n", " detector,\n", " mask=torch.tensor([1., 0.8,])\n", ")\n", "\n", "train_dataset = ReservoirResponseDataset(\n", " mg_dataset,\n", " reservoir_system\n", ")" ] }, { "cell_type": "markdown", "metadata": {}, "source": "## Neural network" }, { "cell_type": "markdown", "metadata": {}, "source": [ "After the reservoir system with a detector, the linear output layer is applied.\n", "The weights of this layer will be trained during the training process.\n", "The resulting neural network is as follows:" ] }, { "cell_type": "code", "execution_count": 47, "metadata": {}, "outputs": [], "source": [ "class Net(torch.nn.Module):\n", " def __init__(self, simulation_parameters: SimulationParameters) -> None:\n", " super().__init__()\n", " self.linear_layer = torch.nn.Linear(\n", " in_features=simulation_parameters.axes_size('W')[0] * simulation_parameters.axes_size('H')[0],\n", " out_features=1\n", " )\n", "\n", " def forward(self, x):\n", " return self.linear_layer(x).squeeze()" ] }, { "cell_type": "code", "execution_count": 48, "metadata": {}, "outputs": [], "source": [ "torch.manual_seed(25)\n", "\n", "net = Net(\n", " sim_params\n", ")" ] }, { "cell_type": "code", "execution_count": 49, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "" ] }, "execution_count": 49, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "with torch.no_grad():\n", " x, y_reservoir, y = train_dataset[40]\n", " plt.plot(y[MG_X_PREHEAT_NUM:], label='(Reservoir + Net) required output')\n", " plt.plot(net(y_reservoir)[MG_X_PREHEAT_NUM:].numpy(force=True), label='(Reservoir + Net) output')\n", "\n", "plt.xlabel('symbol')\n", "plt.legend()" ] }, { "cell_type": "markdown", "metadata": {}, "source": "## Training" }, { "cell_type": "markdown", "metadata": {}, "source": [ "Some default training routines" ] }, { "cell_type": "code", "execution_count": 50, "metadata": {}, "outputs": [], "source": [ "train_dataloader = torch.utils.data.DataLoader(train_dataset, batch_size=1, shuffle=True)\n", "\n", "optimizer = torch.optim.AdamW(\n", " params=net.parameters(),\n", " lr=2e-2,\n", ")\n", "\n", "loss_func = torch.nn.MSELoss()\n", "\n", "epoch_train_losses = []" ] }, { "cell_type": "code", "execution_count": 51, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ " 9%|▉ | 9/100 [00:00<00:02, 41.00it/s]" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch #0 mean loss: 1.4420512914657593\n", "Epoch #10 mean loss: 0.12267999351024628\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ " 27%|██▋ | 27/100 [00:00<00:01, 49.60it/s]" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch #20 mean loss: 0.11480734497308731\n", "Epoch #30 mean loss: 0.10821171849966049\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ " 50%|█████ | 50/100 [00:01<00:00, 50.17it/s]" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch #40 mean loss: 0.10631066560745239\n", "Epoch #50 mean loss: 0.10427819192409515\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ " 68%|██████▊ | 68/100 [00:01<00:00, 50.70it/s]" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch #60 mean loss: 0.10449296981096268\n", "Epoch #70 mean loss: 0.10347328335046768\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ " 90%|█████████ | 90/100 [00:01<00:00, 48.50it/s]" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Epoch #80 mean loss: 0.10367363691329956\n", "Epoch #90 mean loss: 0.1031264141201973\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 100/100 [00:02<00:00, 48.25it/s]\n" ] } ], "source": [ "n_epochs = 100\n", "for epoch in tqdm.tqdm(range(n_epochs)):\n", "\n", " train_losses = []\n", " for x, y, u in train_dataloader:\n", "\n", " u_pred = net(y[0])\n", " loss = loss_func(u[0][MG_X_PREHEAT_NUM:], u_pred[MG_X_PREHEAT_NUM:])\n", "\n", " loss.backward()\n", " optimizer.step()\n", " optimizer.zero_grad()\n", "\n", " train_losses.append(loss.item())\n", "\n", " epoch_train_losses.append(torch.mean(torch.tensor(train_losses)).item())\n", " if epoch % 10 == 0:\n", " print(f'Epoch #{epoch} mean loss:', epoch_train_losses[-1])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Evaluation" ] }, { "cell_type": "code", "execution_count": 52, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "" ] }, "execution_count": 52, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "with torch.no_grad():\n", " x, y_reservoir, y = train_dataset[40]\n", " plt.plot(y[MG_X_PREHEAT_NUM:], label='(Reservoir + Net) required output')\n", " plt.plot(net(y_reservoir)[MG_X_PREHEAT_NUM:].numpy(force=True), label='(Reservoir + Net) output')\n", "\n", "plt.xlabel('symbol')\n", "plt.legend()" ] } ], "metadata": { "kernelspec": { "display_name": ".venv", "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": "3.11.9" } }, "nbformat": 4, "nbformat_minor": 2 }