{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "## Compare VOPy with Optuna on Black-box Vector Optimization\n", "This notebook compares VOPy with the state-of-the-art black-box optimization library called **Optuna**. As a real-world problem, we use a Paal-Knorr reaction for optimization. This problem is implemented in Tu et al, 2022. We will use a variant of PaVeBa (Karagözlü et al, 2023) algorithm for online learning (see [PaVeBaGPOnline](../algorithms.rst#pavebagponline) docs).\n", "\n", "- Ben Tu, Axel Gandy, Nikolas Kantas, Behrang Shafei. [Joint Entropy Search for Multi-objective Bayesian Optimization](https://arxiv.org/pdf/2210.02905.pdf). NeurIPS 2022. [GitHub Repo.](https://github.com/benmltu/JES/)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Before starting, make sure the Optuna is installed. We include it as optional-dependencies under `examples`, see `pyproject.toml` in GitHub." ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [], "source": [ "import warnings\n", "from unittest import mock\n", "\n", "import logging\n", "from importlib import reload\n", "\n", "\n", "import matplotlib.pyplot as plt\n", "\n", "import torch\n", "import numpy as np\n", "from scipy.stats.qmc import Sobol\n", "\n", "import optuna\n", "\n", "from vopy.utils import set_seed\n", "from vopy.order import ConeTheta2DOrder\n", "from vopy.algorithms import PaVeBaGPOnline\n", "from vopy.maximization_problem import FixedPointsProblem\n", "from vopy.utils.evaluate import calculate_epsilonF1_score\n", "\n", "warnings.filterwarnings(\"ignore\")\n", "optuna.logging.set_verbosity(optuna.logging.WARNING)\n", "\n", "# reload(logging)\n", "# logging.basicConfig(level=logging.INFO)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### Loading `PK2` from JES dynamically\n", "The code for JES is not properly packaged, so we can not install it using pip. Therefore, we download the file we are interested in and load it manually." ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [], "source": [ "import importlib, sys, pathlib\n", "from urllib.request import urlopen\n", "\n", "RAW_URL_COMMIT = \"https://raw.githubusercontent.com/benmltu/JES/7ea9856781869c0cd348fb52c7431c40274225e9/jes/benchmarks/chembench.py\"\n", "cache_dir = pathlib.Path(\"tmp/jes/\")\n", "cache_dir.mkdir(parents=True, exist_ok=True)\n", "\n", "try:\n", " with urlopen(RAW_URL_COMMIT) as resp:\n", " source_bytes = resp.read()\n", "except Exception as dl_e:\n", " raise RuntimeError(f\"Failed to download chembench.py from JES repo: {dl_e}\")\n", "\n", "cached_path = cache_dir / f\"chembench.py\"\n", "if not cached_path.exists():\n", " cached_path.write_bytes(source_bytes)\n", "\n", "# Load module from cached file\n", "spec_name = f\"_jes_chembench\"\n", "spec = importlib.util.spec_from_file_location(spec_name, cached_path)\n", "\n", "module = importlib.util.module_from_spec(spec)\n", "sys.modules[spec_name] = module\n", "\n", "spec.loader.exec_module(module)\n", "\n", "PK2 = getattr(module, \"PK2\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Fixing a seed for reproducibility." ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [], "source": [ "set_seed(0)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We will create a problem instance for the PK2 problem from JES module. It follows the Botorch API." ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [], "source": [ "noise_var = 0.0001\n", "noise_std = noise_var ** 0.5 # Only known by the problem.\n", "\n", "problem = PK2(noise_std=noise_std, negate=True) # `negate=True` because we are maximizing." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Let's now create our `Order` object to define our preference. Our Pareto set identification and evaluation will be according to this preference." ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [], "source": [ "order = ConeTheta2DOrder(cone_degree=45)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "As we will work in a fixed cardinality space, we fix the possible design choices (chemical experiment parameters in this case) beforehand." ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [], "source": [ "design_count = 500\n", "\n", "input_bounds = problem._bounds\n", "\n", "# Create designs\n", "num_sobol_samples = design_count\n", "sampler = Sobol(problem.dim, scramble=False)\n", "sampler.fast_forward(np.random.randint(low=1, high=max(2, num_sobol_samples)))\n", "X_np = sampler.random(num_sobol_samples)\n", "\n", "# Scale designs to the problem bounds\n", "for i, (l, u) in enumerate(input_bounds):\n", " X_np[:, i] = X_np[:, i] * (u - l) + l" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### Optimizing with different libraries\n", "We'll now setup and run both frameworks (VOPy and Optuna). As algorithms from VOPy has fixed confidence and Optuna has fixed sampling count, we will first run PaVeBaGPOnline from VOPY and then run Optuna using the same number of samples." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### Set up and run VOPy" ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "PaVeBaGPOnline: 30 samples taken so far.\n", "PaVeBaGPOnline: 40 samples taken so far.\n", "PaVeBaGPOnline: 50 samples taken so far.\n", "PaVeBaGPOnline: 60 samples taken so far.\n", "PaVeBaGPOnline: 70 samples taken so far.\n", "PaVeBaGPOnline: 80 samples taken so far.\n", "PaVeBaGPOnline: 90 samples taken so far.\n", "PaVeBaGPOnline: 100 samples taken so far.\n", "PaVeBaGPOnline: 110 samples taken so far.\n", "PaVeBaGPOnline: 120 samples taken so far.\n", "PaVeBaGPOnline took 128 samples (including initial) to finish.\n" ] } ], "source": [ "paveba_problem = FixedPointsProblem(\n", " in_points=X_np,\n", " out_dim=problem.num_objectives,\n", " objective=lambda x: problem(torch.tensor(x)).numpy(force=True),\n", ")\n", "\n", "# Algorithm parameters\n", "epsilon = 0.01\n", "delta = 0.05\n", "\n", "paveba_alg = PaVeBaGPOnline(\n", " epsilon=epsilon,\n", " delta=delta,\n", " problem=paveba_problem,\n", " order=order,\n", " conf_contraction=128,\n", " initial_sample_cnt=25,\n", " reset_on_retrain=False,\n", ")\n", "\n", "while True:\n", " is_done = paveba_alg.run_one_step()\n", "\n", " if paveba_alg.sample_count % 10 == 0:\n", " print(f\"PaVeBaGPOnline: {paveba_alg.sample_count} samples taken so far.\")\n", "\n", " if is_done:\n", " break\n", "\n", "print(f\"PaVeBaGPOnline took {paveba_alg.sample_count} samples (including initial) to finish.\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### Set up and run Optuna" ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Number of trials by Optuna (equals the sample count of PaVeBaGPOnline): 128\n" ] } ], "source": [ "def optuna_objective(trial):\n", " tau = trial.suggest_float(\"tau\", problem._bounds[0][0], problem._bounds[0][1])\n", " temperature = trial.suggest_float(\"temperature\", problem._bounds[1][0], problem._bounds[1][1])\n", " equivalents = trial.suggest_float(\"equivalents\", problem._bounds[2][0], problem._bounds[2][1])\n", " suggested_point = np.array([tau, temperature, equivalents])\n", " \n", " # We will sample the closest point from the design space.\n", " distances = np.linalg.norm(X_np - suggested_point, axis=1)\n", " closest_index = np.argmin(distances)\n", " inp = torch.tensor(X_np[closest_index]).reshape(1, problem.dim)\n", " \n", " return problem(inp).numpy(force=True).tolist()[0]\n", "\n", "X = X_np.tolist()\n", "X = [tuple(x) for x in X]\n", "\n", "study = optuna.create_study(directions=[\"maximize\", \"maximize\"])\n", "study.optimize(optuna_objective, n_trials=paveba_alg.sample_count, timeout=120)\n", "print(\"Number of trials by Optuna (equals the sample count of PaVeBaGPOnline): \", len(study.trials))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Evaluation\n", "\n", "We will evaluate the frameworks. We can start by computing the ground truth values." ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [], "source": [ "Y_true = problem(torch.tensor(X_np), noise=False).numpy(force=True)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We compute the corresponding Pareto set approximations for both frameworks." ] }, { "cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [], "source": [ "optuna_p_set = set()\n", "for trial in study.best_trials:\n", " trial_pt = np.array((trial.params[\"tau\"], trial.params[\"temperature\"], trial.params[\"equivalents\"]))\n", " closest_index = np.argmin(np.linalg.norm(trial_pt - X_np, axis=1))\n", " optuna_p_set.add(closest_index)\n", "optuna_p_set = list(optuna_p_set)\n", "\n", "vopy_p_set = list(paveba_alg.P)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Results in cone order that we defined:" ] }, { "cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "VOPy epsilon-F1 score: 0.88\n", "Optuna epsilon-F1 score: 0.19\n" ] } ], "source": [ "dataset = mock.Mock()\n", "dataset.out_data = Y_true\n", "true_pareto_inds = order.get_pareto_set(Y_true)\n", "\n", "vopy_epsf1 = calculate_epsilonF1_score(dataset, order, true_pareto_inds, vopy_p_set, epsilon)\n", "optuna_epsf1 = calculate_epsilonF1_score(dataset, order, true_pareto_inds, optuna_p_set, epsilon)\n", "\n", "print(f\"VOPy epsilon-F1 score: {vopy_epsf1:.2f}\")\n", "print(f\"Optuna epsilon-F1 score: {optuna_epsf1:.2f}\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "**As you can see, Optuna can not perform under the vector optimization setting, as they don't encode preferences like VOPy does. While Optuna achieves an epsilon-F1 score of 0.19, VOPy (using PaVeBaGPOnline algorithm) achieves an epsilon-F1 score of 0.88, which is significantly higher.**" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### Visualization of Pareto sets" ] }, { "cell_type": "code", "execution_count": 12, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "" ] }, "execution_count": 12, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "tmp_pareto_fig = order.plot_pareto_set(Y_true)\n", "tmp_pareto_fig.axes[0].scatter(\n", " dataset.out_data[vopy_p_set][:, 0],\n", " dataset.out_data[vopy_p_set][:, 1],\n", " c=\"tab:red\", label=\"VOPy Pareto\", alpha=0.5, marker=\"^\"\n", ")\n", "tmp_pareto_fig.axes[0].scatter(\n", " dataset.out_data[optuna_p_set][:, 0],\n", " dataset.out_data[optuna_p_set][:, 1],\n", " c=\"tab:green\", label=\"Optuna Pareto\", alpha=0.5, marker=\"v\"\n", ")\n", "tmp_pareto_fig.axes[0].legend(loc=\"lower left\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "**We can see that the Pareto set found by vector optimization using VOPy matches highly with the ground truth, whereas the Pareto set found by Optuna is lacking a lot of Pareto optimal points.**" ] } ], "metadata": { "kernelspec": { "display_name": "VOPy", "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": "3.13.3" } }, "nbformat": 4, "nbformat_minor": 2 }