{ "cells": [ { "cell_type": "markdown", "id": "91ae3118", "metadata": {}, "source": [ "# Quantum Optical Reservoir Computing (QORC) with MerLin\n", "\n", "This notebook is a first practical introduction to **quantum optical reservoir computing** inspired by:\n", "\n", "- [**Quantum optical reservoir computing powered by boson sampling**](https://opg.optica.org/opticaq/fulltext.cfm?uri=opticaq-3-3-238), Sakurai et al., (2025)\n", "- [**Photonic Quantum-Accelerated Machine Learning**](https://arxiv.org/abs/2512.08318), Rambach et al., (2025)\n", "\n", "Goals:\n", "\n", "1. Explain the reservoir structure and intuition.\n", "2. Build a first reservoir with `QuantumLayer.simple()`.\n", "3. Build a reservoir with the same fixed Haar-random unitary used **before and after** encoding.\n", "4. Evaluate it and compare against classical baselines." ] }, { "cell_type": "markdown", "id": "5ab38606", "metadata": {}, "source": [ "## 1) Reservoir computing intuition\n", "\n", "![Quantum optical reservoir scheme](../_static/img/reservoir_scheme.png)\n", "\n", "\n", "Reservoir computing separates the model into two parts:\n", "\n", "- **Reservoir (A)**: a fixed nonlinear feature map (not trained).\n", "- **Readout**: a small trainable classical model (often linear).\n", "\n", "If we look at the image above, we could have 3 ways to classify the images:\n", "- Option A: a fixed Boson Sampler applied to a specific input (PCA components or raw inputs, depending on the input size and the experiment). The output of this Boson Sampler is fed to a trainable Linear layer that maps it to the correct output (here, the classes)\n", "- Option B: the image is directly fed to the Linear layer and we operate a linear classification\n", "- Option C (and the case in the 2 papers mentioned): the input to the linear readout is the concatenation of the quantum features and the raw (flattened) input\n", "\n", "NOTE: we can imagine using another kind of readout (MLP, CNN). For the purpose of this tutorial and because it is used in the 2 papers, we use a linear readout.\n", "\n", "Why this is useful:\n", "\n", "- training is simpler and faster (few trainable parameters),\n", "- nonlinear dynamics in the reservoir can lift data into a richer feature space.\n", "\n", "In a photonic quantum setting, the reservoir is a fixed interferometer + encoding + measurement pipeline.\n", "\n", "Before diving in this tutorial, let's import all necessary tools:" ] }, { "cell_type": "code", "execution_count": 1, "id": "dd7559d0", "metadata": {}, "outputs": [], "source": [ "import sys\n", "from pathlib import Path\n", "\n", "# Force the notebook to import the local source tree before any installed wheel.\n", "for candidate in [Path.cwd(), *Path.cwd().parents]:\n", " if (candidate / 'merlin' / 'models' / 'reservoir_classifier.py').exists():\n", " repo_root = candidate\n", " break\n", "else:\n", " repo_root = None\n", "\n", "if repo_root is not None and str(repo_root) not in sys.path:\n", " sys.path.insert(0, str(repo_root))\n", "\n", "import time\n", "import numpy as np\n", "import pandas as pd\n", "import matplotlib.pyplot as plt\n", "\n", "import torch\n", "import torch.nn as nn\n", "\n", "import merlin as ML\n", "import perceval as pcvl\n", "from perceval.components import BS, PS\n", "\n", "from sklearn.datasets import make_moons\n", "from sklearn.model_selection import train_test_split\n", "from sklearn.preprocessing import StandardScaler\n", "from sklearn.linear_model import LogisticRegression\n", "from sklearn.neural_network import MLPClassifier\n", "from sklearn.metrics import accuracy_score, f1_score" ] }, { "cell_type": "markdown", "id": "b1b52f3f", "metadata": {}, "source": [ "## 2) Experimental setup\n", "\n", "### 2.1) Loading a simple dataset" ] }, { "cell_type": "code", "execution_count": 2, "id": "eae95f9e", "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "SEED = 7\n", "np.random.seed(SEED)\n", "torch.manual_seed(SEED)\n", "\n", "\n", "def make_dataset(seed=SEED):\n", " X, y = make_moons(n_samples=320, noise=0.20, random_state=seed)\n", " X_train, X_test, y_train, y_test = train_test_split(\n", " X, y, test_size=0.30, stratify=y, random_state=seed\n", " )\n", "\n", " scaler = StandardScaler()\n", " X_train = scaler.fit_transform(X_train)\n", " X_test = scaler.transform(X_test)\n", "\n", " return (\n", " X_train.astype(np.float32),\n", " X_test.astype(np.float32),\n", " y_train.astype(np.int64),\n", " y_test.astype(np.int64),\n", " )\n", "\n", "\n", "X_train, X_test, y_train, y_test = make_dataset()\n", "\n", "plt.figure(figsize=(5, 4))\n", "plt.scatter(X_train[:, 0], X_train[:, 1], c=y_train, s=20, alpha=0.8)\n", "plt.title(\"Training data (two moons)\")\n", "plt.xlabel(\"x1\")\n", "plt.ylabel(\"x2\")\n", "plt.show()" ] }, { "cell_type": "markdown", "id": "9e74ace3", "metadata": {}, "source": [ "### 2.2) QOR structure used here\n", "\n", "We use the following pattern:\n", "\n", "1. fixed interferometer `U` (mixing stage),\n", "2. data encoding `E(x)` (phase shifts from input features),\n", "3. the **same** fixed interferometer `U` again,\n", "4. photonic measurement gives reservoir features,\n", "5. as mentioned above, we train a classical readout on top.\n", "\n", "This follows the common \"fixed random optical map + trainable readout\" idea used in QOR works." ] }, { "cell_type": "markdown", "id": "0840bdcb", "metadata": {}, "source": [ "### Note on PCA\n", "\n", "In the MNIST experiments from the referenced papers, PCA is often applied before encoding to reduce dimensionality.\n", "\n", "For this notebook we use two-moons (2 features), so PCA is not necessary.\n", "\n", "If you later switch this notebook to image datasets (e.g., MNIST), adding a PCA preprocessing block is recommended for a fair protocol match.\n" ] }, { "cell_type": "markdown", "id": "ffdf763b", "metadata": {}, "source": [ "## 3) Classical baselines\n", "\n", "We train two standard baselines:\n", "\n", "- Logistic Regression\n", "- 1-hidden-layer MLP" ] }, { "cell_type": "code", "execution_count": 3, "id": "dd34fa83", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
modelaccuracyf1train_time_s
0LogReg0.8020830.8080810.009799
1MLP(16)0.9166670.9183670.220464
\n", "
" ], "text/plain": [ " model accuracy f1 train_time_s\n", "0 LogReg 0.802083 0.808081 0.009799\n", "1 MLP(16) 0.916667 0.918367 0.220464" ] }, "execution_count": 3, "metadata": {}, "output_type": "execute_result" } ], "source": [ "def run_classical_baselines(X_train, y_train, X_test, y_test):\n", " results = []\n", "\n", " baselines = {\n", " \"LogReg\": LogisticRegression(max_iter=1000, random_state=SEED),\n", " \"MLP(16)\": MLPClassifier(\n", " hidden_layer_sizes=(16,), max_iter=1200, random_state=SEED\n", " ),\n", " }\n", "\n", " for name, model in baselines.items():\n", " t0 = time.perf_counter()\n", " model.fit(X_train, y_train)\n", " dt = time.perf_counter() - t0\n", "\n", " y_pred = model.predict(X_test)\n", " results.append({\n", " \"model\": name,\n", " \"accuracy\": accuracy_score(y_test, y_pred),\n", " \"f1\": f1_score(y_test, y_pred),\n", " \"train_time_s\": dt,\n", " })\n", "\n", " return results\n", "\n", "\n", "classical_results = run_classical_baselines(X_train, y_train, X_test, y_test)\n", "pd.DataFrame(classical_results)" ] }, { "cell_type": "markdown", "id": "6cb32498", "metadata": {}, "source": [ "## 4) Shared helper: cached reservoir features + trainable linear readout\n", "\n", "In the referenced QOR workflows, the readout input is the concatenation `[x, r(x)]`.\n", "\n", "For efficiency, we compute `r(x)` once, cache it, then train only a linear readout on cached features.\n" ] }, { "cell_type": "code", "execution_count": 4, "id": "b86a7f25", "metadata": {}, "outputs": [], "source": [ "def compute_reservoir_features(reservoir, X, batch_size=128):\n", " reservoir.requires_grad_(False)\n", " reservoir.eval()\n", "\n", " X_tensor = torch.tensor(X, dtype=torch.float32)\n", " chunks = []\n", "\n", " with torch.no_grad():\n", " for start in range(0, len(X_tensor), batch_size):\n", " xb = X_tensor[start : start + batch_size]\n", " rb = reservoir(xb)\n", " chunks.append(rb)\n", "\n", " return torch.cat(chunks, dim=0)\n", "\n", "\n", "def make_concat_features(reservoir, X):\n", " X_tensor = torch.tensor(X, dtype=torch.float32)\n", " R_tensor = compute_reservoir_features(reservoir, X)\n", " return torch.cat([X_tensor, R_tensor], dim=1)\n", "\n", "\n", "def train_linear_readout_from_features(\n", " F_train, y_train, F_test, y_test, epochs=200, lr=0.05\n", "):\n", " ytr = torch.tensor(y_train, dtype=torch.long)\n", " yte = torch.tensor(y_test, dtype=torch.long)\n", " n_classes = int(np.max(y_train)) + 1\n", "\n", " readout = nn.Linear(F_train.shape[1], n_classes)\n", " optimizer = torch.optim.Adam(readout.parameters(), lr=lr)\n", " loss_fn = nn.CrossEntropyLoss()\n", "\n", " # Track training history\n", " history = {\n", " \"train_loss\": [],\n", " \"val_loss\": [],\n", " \"train_acc\": [],\n", " \"val_acc\": [],\n", " }\n", "\n", " t0 = time.perf_counter()\n", " readout.train()\n", " for epoch in range(epochs):\n", " # Training step\n", " optimizer.zero_grad()\n", " logits = readout(F_train)\n", " loss = loss_fn(logits, ytr)\n", " loss.backward()\n", " optimizer.step()\n", "\n", " # Track metrics every epoch\n", " readout.eval()\n", " with torch.no_grad():\n", " # Train metrics\n", " train_logits = readout(F_train)\n", " train_loss = loss_fn(train_logits, ytr).item()\n", " train_pred = train_logits.argmax(dim=1).cpu().numpy()\n", " train_acc = accuracy_score(y_train, train_pred)\n", "\n", " # Val metrics\n", " val_logits = readout(F_test)\n", " val_loss = loss_fn(val_logits, yte).item()\n", " val_pred = val_logits.argmax(dim=1).cpu().numpy()\n", " val_acc = accuracy_score(y_test, val_pred)\n", "\n", " history[\"train_loss\"].append(train_loss)\n", " history[\"val_loss\"].append(val_loss)\n", " history[\"train_acc\"].append(train_acc)\n", " history[\"val_acc\"].append(val_acc)\n", " readout.train()\n", "\n", " dt = time.perf_counter() - t0\n", "\n", " readout.eval()\n", " with torch.no_grad():\n", " y_pred = readout(F_test).argmax(dim=1).cpu().numpy()\n", "\n", " return {\n", " \"accuracy\": accuracy_score(y_test, y_pred),\n", " \"f1\": f1_score(y_test, y_pred),\n", " \"train_time_s\": dt,\n", " \"history\": history,\n", " }" ] }, { "cell_type": "markdown", "id": "deb76f95", "metadata": {}, "source": [ "## 5) First reservoir with `QuantumLayer.simple()`\n", "\n", "This gives a quick fixed photonic feature map. We explicitly choose the computation space in the current API, then learn only the readout." ] }, { "cell_type": "code", "execution_count": 5, "id": "ec719d32", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "/var/folders/g5/xyqwmpys7177hgy25td4j6740000gn/T/ipykernel_89624/375005998.py:1: DeprecationWarning: Parameter 'computation_space' is deprecated. The 'computation_space' keyword is deprecated; move it into MeasurementStrategy.probs(computation_space).\n", " simple_reservoir = ML.QuantumLayer.simple(\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Results: Accuracy 0.7812, F1 0.7921, Train time 0.18s\n" ] } ], "source": [ "simple_reservoir = ML.QuantumLayer.simple(\n", " input_size=2,\n", " computation_space=ML.ComputationSpace.UNBUNCHED,\n", ")\n", "\n", "F_train_simple = make_concat_features(simple_reservoir, X_train)\n", "F_test_simple = make_concat_features(simple_reservoir, X_test)\n", "\n", "simple_metrics = train_linear_readout_from_features(\n", " F_train_simple, y_train, F_test_simple, y_test, epochs=240, lr=0.01\n", ")\n", "\n", "print(\n", " f\"Results: Accuracy {simple_metrics['accuracy']:.4f}, F1 {simple_metrics['f1']:.4f}, Train time {simple_metrics['train_time_s']:.2f}s\"\n", ")" ] }, { "cell_type": "markdown", "id": "4e626472", "metadata": {}, "source": [ "## 6) ReservoirClassifier directly on the 2D input\n", "\n", "Instead of manually building a custom `QuantumLayer`, we can use the higher-level `ReservoirClassifier` API.\n", "\n", "For the two-moons dataset we keep the original 2 features and feed them directly to `ReservoirClassifier`. Without a reduction step, the reservoir therefore stays a true 2-mode photonic model." ] }, { "cell_type": "code", "execution_count": 6, "id": "96ae300d", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Input width: 2\n", "Reservoir modes: 2, quantum output size: 1, readout input width: 3\n", "Results: Accuracy 0.8021, F1 0.8081, Train time 0.17s\n" ] } ], "source": [ "def make_reservoir_classifier_inputs(X_train, X_test):\n", " return X_train.astype(np.float32), X_test.astype(np.float32)\n", "\n", "\n", "def train_reservoir_classifier(model, X_train, y_train, X_test, y_test, epochs=240, lr=0.01):\n", " model.fit_reservoir(X_train)\n", " F_train, ytr = model.make_dataset(X_train, y_train).tensors\n", " F_test, yte = model.make_dataset(X_test, y_test).tensors\n", "\n", " optimizer = torch.optim.Adam(model.parameters(), lr=lr)\n", " loss_fn = nn.CrossEntropyLoss()\n", "\n", " history = {\n", " 'train_loss': [],\n", " 'val_loss': [],\n", " 'train_acc': [],\n", " 'val_acc': [],\n", " }\n", "\n", " t0 = time.perf_counter()\n", " for epoch in range(epochs):\n", " model.train()\n", " optimizer.zero_grad()\n", " logits = model(F_train)\n", " loss = loss_fn(logits, ytr)\n", " loss.backward()\n", " optimizer.step()\n", "\n", " model.eval()\n", " with torch.no_grad():\n", " train_logits = model(F_train)\n", " val_logits = model(F_test)\n", " history['train_loss'].append(loss_fn(train_logits, ytr).item())\n", " history['val_loss'].append(loss_fn(val_logits, yte).item())\n", " history['train_acc'].append(accuracy_score(y_train, train_logits.argmax(dim=1).cpu().numpy()))\n", " history['val_acc'].append(accuracy_score(y_test, val_logits.argmax(dim=1).cpu().numpy()))\n", "\n", " dt = time.perf_counter() - t0\n", " y_pred = model.predict(X_test).argmax(dim=1).cpu().numpy()\n", "\n", " return {\n", " 'accuracy': accuracy_score(y_test, y_pred),\n", " 'f1': f1_score(y_test, y_pred),\n", " 'train_time_s': dt,\n", " 'history': history,\n", " }\n", "\n", "\n", "X_train_reservoir, X_test_reservoir = make_reservoir_classifier_inputs(X_train, X_test)\n", "\n", "reservoir_classifier = ML.ReservoirClassifier(\n", " in_features=X_train_reservoir.shape[1],\n", " out_features=2,\n", " n_photons=2,\n", " concatenate=True,\n", " cache=True,\n", " seed=SEED,\n", " dtype=torch.float32,\n", ")\n", "reservoir_classifier.layer.measurement_strategy = ML.MeasurementStrategy.probs(\n", " computation_space=ML.ComputationSpace.UNBUNCHED,\n", ")\n", "\n", "reservoir_classifier_metrics = train_reservoir_classifier(\n", " reservoir_classifier,\n", " X_train_reservoir,\n", " y_train,\n", " X_test_reservoir,\n", " y_test,\n", " epochs=320,\n", " lr=0.02,\n", ")\n", "\n", "print(f\"Input width: {X_train_reservoir.shape[1]}\")\n", "print(\n", " f\"Reservoir modes: {reservoir_classifier.layer.n_modes}, \"\n", " f\"quantum output size: {reservoir_classifier.layer.output_size}, \"\n", " f\"readout input width: {reservoir_classifier.readout.in_features}\"\n", ")\n", "print(\n", " f\"Results: Accuracy {reservoir_classifier_metrics['accuracy']:.4f}, \"\n", " f\"F1 {reservoir_classifier_metrics['f1']:.4f}, \"\n", " f\"Train time {reservoir_classifier_metrics['train_time_s']:.2f}s\"\n", ")" ] }, { "cell_type": "markdown", "id": "5d01e28f", "metadata": {}, "source": [ "## 7) Compare all models" ] }, { "cell_type": "code", "execution_count": 7, "id": "fb695bd6", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
modelaccuracyf1train_time_shistory
1MLP(16)0.9166670.9183670.220464NaN
0LogReg0.8020830.8080810.009799NaN
3ReservoirClassifier (2 features)0.8020830.8080810.174888{'train_loss': [0.6682752966880798, 0.64041745...
2MerLin simple0.7812500.7920790.181588{'train_loss': [0.8916221261024475, 0.87204307...
\n", "
" ], "text/plain": [ " model accuracy f1 train_time_s \\\n", "1 MLP(16) 0.916667 0.918367 0.220464 \n", "0 LogReg 0.802083 0.808081 0.009799 \n", "3 ReservoirClassifier (2 features) 0.802083 0.808081 0.174888 \n", "2 MerLin simple 0.781250 0.792079 0.181588 \n", "\n", " history \n", "1 NaN \n", "0 NaN \n", "3 {'train_loss': [0.6682752966880798, 0.64041745... \n", "2 {'train_loss': [0.8916221261024475, 0.87204307... " ] }, "execution_count": 7, "metadata": {}, "output_type": "execute_result" } ], "source": [ "all_results = []\n", "all_results.extend(classical_results)\n", "\n", "all_results.append({\n", " \"model\": \"MerLin simple\",\n", " **simple_metrics,\n", "})\n", "\n", "all_results.append({\n", " 'model': 'ReservoirClassifier (2 features)',\n", " **reservoir_classifier_metrics,\n", "})\n", "\n", "results_df = pd.DataFrame(all_results).sort_values(by=\"accuracy\", ascending=False)\n", "results_df" ] }, { "cell_type": "code", "execution_count": 8, "id": "12089a0a", "metadata": {}, "outputs": [], "source": [ "def plot_training_curves(history, title):\n", " \"\"\"Plot training and validation loss/accuracy curves.\"\"\"\n", " fig, axes = plt.subplots(1, 2, figsize=(12, 4))\n", "\n", " # Plot loss\n", " axes[0].plot(history[\"train_loss\"], label=\"Train Loss\", linewidth=2)\n", " axes[0].plot(history[\"val_loss\"], label=\"Val Loss\", linewidth=2)\n", " axes[0].set_xlabel(\"Epoch\")\n", " axes[0].set_ylabel(\"Loss\")\n", " axes[0].set_title(f\"{title} - Loss\")\n", " axes[0].legend()\n", " axes[0].grid(True, alpha=0.3)\n", "\n", " # Plot accuracy\n", " axes[1].plot(history[\"train_acc\"], label=\"Train Accuracy\", linewidth=2)\n", " axes[1].plot(history[\"val_acc\"], label=\"Val Accuracy\", linewidth=2)\n", " axes[1].set_xlabel(\"Epoch\")\n", " axes[1].set_ylabel(\"Accuracy\")\n", " axes[1].set_title(f\"{title} - Accuracy\")\n", " axes[1].legend()\n", " axes[1].grid(True, alpha=0.3)\n", "\n", " plt.tight_layout()\n", " plt.show()" ] }, { "cell_type": "code", "execution_count": 9, "id": "b088b1b7", "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "plot_training_curves(simple_metrics[\"history\"], \"MerLin Simple Reservoir\")" ] }, { "cell_type": "code", "execution_count": 10, "id": "d1ca216c", "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "plot_training_curves(reservoir_classifier_metrics['history'], 'ReservoirClassifier (2 features)')\n" ] }, { "cell_type": "markdown", "id": "96b20875", "metadata": {}, "source": [ "### Analysis\n", "\n", "**Accuracy Performance:**\n", "\n", "Looking at the results above, we compare the models on the Moon dataset:\n", "\n", "- **LogReg**: Linear baseline with relatively low accuracy due to the non-linearity of the problem.\n", "- **MLP(16)**: Classical neural network with one hidden layer achieves good accuracy.\n", "- **MerLin simple**: Fixed photonic feature map provides non-linear features for the linear readout.\n", "- **ReservoirClassifier (2 features)**: the built-in frozen photonic reservoir is applied directly to the original two-moons coordinates, with a trainable readout on top.\n", "\n", "**Overfitting Analysis:**\n", "\n", "From the training curves, we observe:\n", "\n", "- **Simple Reservoir**: Shows clear **overfitting** — the training loss continues to decrease while validation loss increases or plateaus. The gap between red and blue lines in the accuracy plot widens, indicating the model is memorizing training data rather than generalizing.\n", "\n", "- **ReservoirClassifier**: typically gives smoother optimization because the reservoir features are standardized internally and the readout is trained on a fixed cached feature tensor.\n", "\n", "This demonstrates how the higher-level `ReservoirClassifier` API can replace custom circuit plumbing while keeping the same frozen-reservoir workflow.\n", "\n", "**Training Time:**\n", "\n", "- Classical methods (LogReg, MLP) train in **milliseconds**, as they operate on raw inputs.\n", "- Quantum reservoirs require **computing reservoir features** (photonic simulation) for each input, making them slower. The `ReservoirClassifier` and simple reservoir examples are dominated by readout training once the features are cached.\n", "- In a practical photonic setting, the reservoir computation would be parallelized on dedicated hardware, closing this gap.\n", "\n", "**Key Takeaways:**\n", "\n", "1. The higher-level `ReservoirClassifier` API is enough for a complete frozen-reservoir workflow on this task.\n", "2. On this toy dataset, skipping PCA keeps the example aligned with the original 2-feature input and makes the reservoir structure easier to interpret.\n", "3. For real quantum hardware, the speedup would be more pronounced, as photonic operations execute in constant time regardless of feature space dimensionality." ] }, { "cell_type": "markdown", "id": "318a22da", "metadata": {}, "source": [ "## 8) Sensitivity to the number of photons in a 2-mode reservoir\n", "\n", "In this example we keep the raw 2-feature input and leave the reservoir at 2 modes. A meaningful knob to vary is therefore the number of photons in the input state.\n", "\n", "If you manually increase `layer.n_modes`, only the first encoded input modes receive phase shifters, while the extra modes remain available for the interferometer dynamics.\n" ] }, { "cell_type": "code", "execution_count": 16, "id": "bcb07ac7", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Circuit with 3 modes and 1 photons:\n", "Circuit with 3 modes and 2 photons:\n" ] }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "\n", "Accuracy by photon count:\n", " 1 photons: 0.8021\n", " 2 photons: 0.8125\n" ] } ], "source": [ "SEED = 16\n", "all_photons = [1, 2]\n", "accuracies = {}\n", "for n_photons in all_photons:\n", " np.random.seed(SEED)\n", " torch.manual_seed(SEED)\n", "\n", " X_train_reservoir, X_test_reservoir = make_reservoir_classifier_inputs(X_train, X_test)\n", "\n", " model = ML.ReservoirClassifier(\n", " in_features=X_train_reservoir.shape[1],\n", " out_features=2,\n", " n_photons=n_photons,\n", " concatenate=True,\n", " cache=True,\n", " seed=SEED,\n", " dtype=torch.float32,\n", " )\n", " model.layer.measurement_strategy = ML.MeasurementStrategy.probs(\n", " computation_space=ML.ComputationSpace.UNBUNCHED,\n", " )\n", " model.layer.n_modes = 4\n", " print(f\"Circuit with {model.layer.n_modes} modes and {n_photons} photons:\")\n", " pcvl.pdisplay(model.layer.circuit)\n", "\n", " metrics = train_reservoir_classifier(\n", " model,\n", " X_train_reservoir,\n", " y_train,\n", " X_test_reservoir,\n", " y_test,\n", " epochs=240,\n", " lr=0.02,\n", " )\n", " accuracies[str(n_photons)] = metrics['accuracy']\n", "\n", "photons = [int(k) for k in accuracies.keys()]\n", "accs = [accuracies[k] for k in sorted(accuracies.keys(), key=lambda x: int(x))]\n", "\n", "plt.figure(figsize=(8, 5))\n", "plt.plot(photons, accs, marker='o', linewidth=2, markersize=8, color='blue', label='ReservoirClassifier Accuracy')\n", "plt.scatter(photons, accs, s=100, color='blue', alpha=0.6)\n", "plt.xlabel('Number of Photons', fontsize=12)\n", "plt.ylabel('Test Accuracy', fontsize=12)\n", "plt.title('ReservoirClassifier: Accuracy vs Number of Photons', fontsize=14)\n", "plt.grid(True, alpha=0.3)\n", "plt.xticks(photons)\n", "plt.ylim([0.5, 1.0])\n", "plt.legend()\n", "plt.tight_layout()\n", "plt.show()\n", "\n", "print(f\"\\nAccuracy by photon count:\")\n", "for n_photons in sorted(photons):\n", " print(f\" {n_photons} photons: {accuracies[str(n_photons)]:.4f}\")\n" ] }, { "cell_type": "markdown", "id": "65216cb6", "metadata": {}, "source": [ "**Interpretation**: the performance increases with the size of the Fock space." ] }, { "cell_type": "markdown", "id": "59ded8e5", "metadata": {}, "source": [ "## 9) Interpretation and next step\n", "\n", "What to look for:\n", "\n", "- If `ReservoirClassifier` beats the simple reservoir, the fixed haar-random interferometer likely produced a better photonic feature map directly from the 2D input.\n", "- If a classical MLP still wins, that is normal on small toy data; the goal here is to validate the frozen-reservoir pipeline and evaluation protocol.\n", "\n", "We invite you to:\n", "\n", "1. repeat runs over multiple random seeds,\n", "2. observe the sensitivity to `n_photons`, the measurement strategy, and the random seed,\n", "3. replace synthetic data with a target dataset from your project,\n", "4. richer readouts (ridge/logistic) and regularization." ] } ], "metadata": { "kernelspec": { "display_name": "venv-0.4", "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": "3.12.7" } }, "nbformat": 4, "nbformat_minor": 5 }