01 - Qiskit Quickstart: Classification with Sampler and NoiseProfile¶

This notebook shows how to:

• Build a small parameterized Qiskit circuit.
• Wrap a Sampler with wrap_qiskit_sampler.
• Use ShotBootstrap for predictive distributions.
• Run NoiseProfile to study stability vs. shots.

In [1]:

import matplotlib.pyplot as plt
import numpy as np
from qiskit.circuit import Parameter, QuantumCircuit
from qiskit.primitives import Sampler

from quantumuq import NoiseProfile, ShotBootstrap, ece, nll, wrap_qiskit_sampler
from quantumuq.datasets.toy import make_moons

rng = np.random.default_rng(0)
dataset = make_moons(n_samples=200, noise=0.1, random_state=0)
X, y = dataset.X, dataset.y

perm = rng.permutation(len(X))
train_idx, test_idx = perm[:150], perm[150:]
X_train, y_train = X[train_idx], y[train_idx]
X_test, y_test = X[test_idx], y[test_idx]

plt.figure(figsize=(4, 3))
plt.scatter(X_train[:, 0], X_train[:, 1], c=y_train, cmap="coolwarm", s=10)
plt.title("Two-moons training data")
plt.show()

import matplotlib.pyplot as plt import numpy as np from qiskit.circuit import Parameter, QuantumCircuit from qiskit.primitives import Sampler from quantumuq import NoiseProfile, ShotBootstrap, ece, nll, wrap_qiskit_sampler from quantumuq.datasets.toy import make_moons rng = np.random.default_rng(0) dataset = make_moons(n_samples=200, noise=0.1, random_state=0) X, y = dataset.X, dataset.y perm = rng.permutation(len(X)) train_idx, test_idx = perm[:150], perm[150:] X_train, y_train = X[train_idx], y[train_idx] X_test, y_test = X[test_idx], y[test_idx] plt.figure(figsize=(4, 3)) plt.scatter(X_train[:, 0], X_train[:, 1], c=y_train, cmap="coolwarm", s=10) plt.title("Two-moons training data") plt.show()

In [2]:

# Build a simple 1-qubit circuit with a single parameter.
theta = Parameter("theta")
qc = QuantumCircuit(1)
qc.ry(theta, 0)
qc.measure_all()

def feature_map(X: np.ndarray):
    X_arr = np.asarray(X)
    if X_arr.ndim == 1:
        X_arr = X_arr[0:1]
    # Map first feature to rotation.
    return [[float(x[0])] for x in X_arr]

sampler = Sampler()

predictor = wrap_qiskit_sampler(
    sampler=sampler,
    circuit=qc,
    task="classification",
    n_classes=2,
    feature_map=feature_map,
)

uq = ShotBootstrap(n_samples=8, shots=1000, seed=0)
uq_model = predictor.with_uq(uq)

dist = uq_model.predict_dist(X_test)
probs = dist.mean
y_pred = probs.argmax(axis=1)

test_nll = nll(y_test, probs)
test_ece = ece(y_test, probs, n_bins=10)
print(f"Test NLL: {test_nll:.3f}, ECE: {test_ece:.3f}")

# Build a simple 1-qubit circuit with a single parameter. theta = Parameter("theta") qc = QuantumCircuit(1) qc.ry(theta, 0) qc.measure_all() def feature_map(X: np.ndarray): X_arr = np.asarray(X) if X_arr.ndim == 1: X_arr = X_arr[0:1] # Map first feature to rotation. return [[float(x[0])] for x in X_arr] sampler = Sampler() predictor = wrap_qiskit_sampler( sampler=sampler, circuit=qc, task="classification", n_classes=2, feature_map=feature_map, ) uq = ShotBootstrap(n_samples=8, shots=1000, seed=0) uq_model = predictor.with_uq(uq) dist = uq_model.predict_dist(X_test) probs = dist.mean y_pred = probs.argmax(axis=1) test_nll = nll(y_test, probs) test_ece = ece(y_test, probs, n_bins=10) print(f"Test NLL: {test_nll:.3f}, ECE: {test_ece:.3f}")

Test NLL: 1.084, ECE: 0.271

/var/folders/rh/1c0lrj_x0x956417g86lc4ph0000gn/T/ipykernel_16656/2109316916.py:14: DeprecationWarning: The class ``qiskit.primitives.sampler.Sampler`` is deprecated as of qiskit 1.2. It will be removed no earlier than 3 months after the release date. All implementations of the `BaseSamplerV1` interface have been deprecated in favor of their V2 counterparts. The V2 alternative for the `Sampler` class is `StatevectorSampler`.
  sampler = Sampler()

In [3]:

# NoiseProfile: shot sweep
profile = NoiseProfile(sweep_shots=[100, 500, 1000], n_repeats=4)
results = profile(predictor, X_test[:40])

shots_list = sorted(results.keys())
mean_entropies = [results[s]["mean_entropy"].mean() for s in shots_list]
std_entropies = [results[s]["std_entropy"].mean() for s in shots_list]

plt.figure(figsize=(5, 3))
plt.errorbar(shots_list, mean_entropies, yerr=std_entropies, fmt="o-")
plt.xlabel("Shots")
plt.ylabel("Predictive entropy (mean ± std)")
plt.title("Uncertainty vs shots (Qiskit Sampler)")
plt.grid(True, alpha=0.3)
plt.show()

# NoiseProfile: shot sweep profile = NoiseProfile(sweep_shots=[100, 500, 1000], n_repeats=4) results = profile(predictor, X_test[:40]) shots_list = sorted(results.keys()) mean_entropies = [results[s]["mean_entropy"].mean() for s in shots_list] std_entropies = [results[s]["std_entropy"].mean() for s in shots_list] plt.figure(figsize=(5, 3)) plt.errorbar(shots_list, mean_entropies, yerr=std_entropies, fmt="o-") plt.xlabel("Shots") plt.ylabel("Predictive entropy (mean ± std)") plt.title("Uncertainty vs shots (Qiskit Sampler)") plt.grid(True, alpha=0.3) plt.show()

Interpreting the Qiskit quickstart¶

• The Sampler primitive returns quasi-distributions over bitstrings.
• wrap_qiskit_sampler converts these to class probabilities with a default bitstring-to-class mapping (integer value modulo n_classes).
• ShotBootstrap resamples the circuit multiple times to approximate the predictive distribution over outputs.
• NoiseProfile sweeps over different shot counts and summarizes how probabilities and entropies vary with measurement statistics.