Question 1. What is the 𝜒2 test?
- (a) It is a method to avoid overfitting
- (b) It is a statistical method of feature selection
- (c) It is a feature selection method based on training
- (d) It is a statistical measure of a classifier’s accuracy
- (e) It is an empirical method for measuring the accuracy of a supervised classifier
Question 2. Let D = {xi} be a dataset of objects xi. Let Dk = {kxi} be the new dataset obtained by multiplying each coordinate of the objects xi by a constant k. Which of the following statements is incorrect:
- (a) If D is unbalanced, then kD is also unbalanced
- (b) Applying the Nearest Mean Classifier (NMC) to D and kD would yield the same accuracy
- (c) Applying the Quantum Centroid Classifier (QCC) to D and kD would yield the same accuracy
- (d) Applying the Quantum Centroid Classifier (QCC) to D and kD would not yield the same accuracy
- (e) None of the above
Question 3. Quantum State Discrimination (QSD):
- (a) has no classical analogue because it is based on the concept of quantum measurement
- (b) is the quantum counterpart of Classical State Discrimination
- (c) is based on the fact that, after interaction, two physical systems can share a certain property even at an arbitrary distance from each other
- (d) is based on the ability to distinguish between microscopic and macroscopic phenomena
- (e) is the formal representation of Heisenberg’s uncertainty principle
Question 4. Let D = {xi} be a dataset with cardinality n. Let DQ be the dataset obtained by encoding each element xi of the dataset D in a density operator ρxi. Finally, let ⊗DQ be the dataset obtained by a tensor copy of each element ρxi by itself (ρxi ⊗ ρxi ). Let A be the centroid of D, B the centroid of DQ, and C the centroid of ⊗DQ. We can say that:
- (a) C = A ⊗ A
- (b) C = B ⊗ B
- (c) B = C ⊗ C
- (d) B = A ⊗ A
- (e) None of the above
Question 5. The Pretty Good Classifier (PGM) has been introduced:
(a) as a version of the Helstrom Quantum Classifier able to run on a Quantum Computer
(b) as a natural multi-class extension of the Helstrom Quantum Classifier
(c) as a quantum translation of the One Versus One multi-class classifier
(d) as a sub-optimal quantum-inspired classifier that allows for multi-class classification
(e) as a multi-class extension of the Quantum Centroid Classifier
Would greatly appreciate if you answer these questions above. Thank you.