QML Foundations: Why Quantum Machine Learning?

Beginner Quantum Machine Learning

Created by Pavel · 11.03.2026 at 14:32 UTC

Quantum machine learning is almost always hybrid: classical code optimizes parameters while a small quantum subroutine produces features or expectations. That split is not a compromise for beginners—it is how near-term devices participate in learning.

NISQ hardware is noisy; claims need classical baselines and honest resource counts. Orientation: [1], [2].

Sources

University approvals: 0

Tasks

Question 1

Which statement best matches the Lecture 1 view of QML in practice?

Hint

Think NISQ-era workflows.

QML fully replaces all classical optimization and modeling.

QML is typically hybrid: classical optimization + quantum circuit components.

QML is useful only after fault-tolerant quantum computers exist.

QML has no need for baseline comparisons.

Question 2

Why is a strong classical baseline mandatory in QML experiments?

Hint

Focus on scientific comparison.

to verify whether quantum methods add real value

none of the above

all of the above

insufficient information

Question 3

Which is NOT a core evaluation dimension for QML models?

Hint

One option is clearly irrelevant to model quality.

Prediction quality

Resource cost (shots/depth/hardware)

Reproducibility and variance

Font style used in slides

Question 4

What does NISQ stand for?

Hint

Noisy + intermediate scale.

noisy intermediate-scale quantum

near-infinite scalable quantum

non-interacting sampled qubits

new integrated scalable queue

Question 5

Qiskit and PennyLane are best described as:

Hint

They are software ecosystems for quantum workflows.

Competing operating systems

Quantum/Hybrid software frameworks

Only data visualization libraries

Only cryptography toolkits

Question 6

A baseline model has accuracy 0.80 and your QML model has 0.84. What is the relative improvement percentage?

Hint

Compute (0.84 - 0.80) / 0.80.

10%

Question 7

A QML model outperforms a weak baseline but matches a strong tuned classical baseline. What is the most defensible conclusion?

Hint

Scientific claims depend on strong comparator quality.

Quantum advantage is proven because any improvement over a baseline is enough

No clear advantage is established until comparison against strong baselines favors QML

Classical baselines are irrelevant in QML

The result is invalid because QML was used

Question 8

Which experimental design best supports a reliable QML study?

Hint

Think reproducibility and fairness.

Single random split and one training run

Multiple seeds, fixed protocol, and resource-aware classical comparison

Only reporting the best run

Skipping uncertainty reporting to simplify conclusions

Question 9

Implement relative_improvement(baseline: float, model: float) -> float returning (model - baseline) / baseline. Raise ValueError if baseline == 0.

Hint

Guard zero baseline.

def relative_improvement(baseline: float, model: float) -> float:
    # TODO
    pass

Starter code is prefilled; replace TODO blocks with your solution.

Runtime output (stdout/stderr)

2 test cases will be used for grading

Run checks runtime behavior only. Final correctness is evaluated when you submit.

Card Info

Topic: Quantum Machine Learning
Difficulty: Beginner
Completed: 0 users

Creator

Pavel