Experimental Results for Quantum (Enhanced) Support Vector Machine (QSVM) with Quantum Kernel Training (QKT)/Quantum Kernel Alignment (QKA) on IRIS Dataset
Preface
Experimental Results
- The experimental results will be available soon...
Local Optimizers
ADAptive Moment (ADAM)
- The experimental results...
No. |
MaxIter¥ |
Tol§ |
LR† |
Accuracy‡ |
0 |
20 |
1e-08 |
1e-06 |
0.95556 |
1 |
20 |
1e-08 |
5e-06 |
0.95556 |
2 |
20 |
1e-08 |
1e-05 |
0.95556 |
3 |
20 |
1e-08 |
5e-05 |
0.95556 |
4 |
20 |
1e-08 |
0.0001 |
0.95556 |
5 |
20 |
1e-08 |
0.0005 |
0.95556 |
6 |
20 |
1e-08 |
0.001 |
0.95556 |
7 |
20 |
1e-08 |
0.005 |
0.95556 |
8 |
20 |
1e-08 |
0.01 |
0.95556 |
9 |
20 |
1e-08 |
0.05 |
0.95556 |
10 |
20 |
1e-08 |
0.1 |
0.95556 |
11 |
20 |
1e-08 |
0.5 |
0.95556 |
12 |
20 |
1e-08 |
1.0 |
1.00000 |
13 |
20 |
1e-07 |
1e-06 |
0.95556 |
14 |
20 |
1e-07 |
5e-06 |
0.95556 |
15 |
20 |
1e-07 |
1e-05 |
0.95556 |
16 |
20 |
1e-07 |
5e-05 |
0.95556 |
17 |
20 |
1e-07 |
0.0001 |
0.95556 |
18 |
20 |
1e-07 |
0.0005 |
0.95556 |
19 |
20 |
1e-07 |
0.001 |
0.95556 |
20 |
20 |
1e-07 |
0.005 |
0.95556 |
21 |
20 |
1e-07 |
0.01 |
0.95556 |
22 |
20 |
1e-07 |
0.05 |
0.95556 |
23 |
20 |
1e-07 |
0.1 |
0.95556 |
24 |
20 |
1e-07 |
0.5 |
0.95556 |
25 |
20 |
1e-07 |
1.0 |
1.00000 |
26 |
20 |
1e-06 |
1e-06 |
0.95556 |
27 |
20 |
1e-06 |
5e-06 |
0.95556 |
28 |
20 |
1e-06 |
1e-05 |
0.95556 |
29 |
20 |
1e-06 |
5e-05 |
0.95556 |
30 |
20 |
1e-06 |
0.0001 |
0.95556 |
31 |
20 |
1e-06 |
0.0005 |
0.95556 |
32 |
20 |
1e-06 |
0.001 |
0.95556 |
33 |
20 |
1e-06 |
0.005 |
0.95556 |
34 |
20 |
1e-06 |
0.01 |
0.95556 |
35 |
20 |
1e-06 |
0.05 |
0.95556 |
36 |
20 |
1e-06 |
0.1 |
0.95556 |
37 |
20 |
1e-06 |
0.5 |
0.95556 |
38 |
20 |
1e-06 |
1.0 |
1.00000 |
39 |
20 |
1e-05 |
1e-06 |
0.95556 |
40 |
20 |
1e-05 |
5e-06 |
0.95556 |
41 |
20 |
1e-05 |
1e-05 |
0.95556 |
42 |
20 |
1e-05 |
5e-05 |
0.95556 |
43 |
20 |
1e-05 |
0.0001 |
0.95556 |
44 |
20 |
1e-05 |
0.0005 |
0.95556 |
45 |
20 |
1e-05 |
0.001 |
0.95556 |
46 |
20 |
1e-05 |
0.005 |
0.95556 |
47 |
20 |
1e-05 |
0.01 |
0.95556 |
48 |
20 |
1e-05 |
0.05 |
0.95556 |
49 |
20 |
1e-05 |
0.1 |
0.95556 |
50 |
20 |
1e-05 |
0.5 |
0.95556 |
51 |
20 |
1e-05 |
1.0 |
1.00000 |
52 |
20 |
0.0001 |
1e-06 |
0.95556 |
53 |
20 |
0.0001 |
5e-06 |
0.95556 |
54 |
20 |
0.0001 |
1e-05 |
0.95556 |
55 |
20 |
0.0001 |
5e-05 |
0.95556 |
56 |
20 |
0.0001 |
0.0001 |
0.95556 |
57 |
20 |
0.0001 |
0.0005 |
0.95556 |
58 |
20 |
0.0001 |
0.001 |
0.95556 |
59 |
20 |
0.0001 |
0.005 |
0.95556 |
60 |
20 |
0.0001 |
0.01 |
0.95556 |
61 |
20 |
0.0001 |
0.05 |
0.95556 |
62 |
20 |
0.0001 |
0.1 |
0.95556 |
63 |
20 |
0.0001 |
0.5 |
0.95556 |
64 |
20 |
0.0001 |
1.0 |
1.00000 |
Note(s):
- For these results, it was used the Quantum Support Vector Classifier (QSVC) from Qiskit library, namely through the Qiskit Machine Learning module. For more information, see the following link:
- Additionally, was also used the Quantum Kernel Trainer for the Quantum Kernel Training (QKT)/Quantum Kernel Alignment (QKA) complementar task, namely through the Qiskit Machine Learning module. For more information, see the following link:
- For the optimization algorithm chosen for the Quantum Kernel Trainer, it was used the ADAptive Moment (ADAM) optimizer. For more information, see the following links:
Footnote(s):
- ¥ - This hyperparameter represents the Maximum Number of Iterations of the optimizer.
- § - This hyperparameter represents the Tolerance Error of the optimizer.
- † - This hyperparameter represents the Learning Rate of the optimizer.
- ‡ - The higher Accuracy values are highlighted in green and the lowest Accuracy values are highlighted in red.
AccuMulated Squared GRADient (AMSGRAD)
- The experimental results will be available soon...
Conjugate Gradient (CG)
- The experimental results will be available soon...
Constrained Optimization By Linear Approximation (COBYLA)
- The experimental results will be available soon...
Limited-memory - Broyden-Fletcher-Goldfarb-Shanno - Bound (L-BFGS-B)
- The experimental results will be available soon...
Gaussian-Smoothed Line Search (GSLS)
- The experimental results will be available soon...
Gradient Descent (GD)
- The experimental results will be available soon...
Nakanishi-Fujii-Todo (NFT)
- The experimental results will be available soon...
Nelder Mead (NM)
- The experimental results will be available soon...
Parallelized - limited-memory - Broyden-Fletcher-Goldfarb-Shanno (P-BFGS)
- The experimental results will be available soon...
Powell
- The experimental results will be available soon...
Sequential Least SQuares Programming (SLSQP)
- The experimental results will be available soon...
Simultaneous Perturbation Stochastic Approximation (SPSA)
- The experimental results will be available soon...
Quantum Natural - Simultaneous Perturbation Stochastic Approximation (QNSPSA)
- The experimental results will be available soon...
Truncated Newton Conjugate (TNC)
- The experimental results will be available soon...
(Continuous) Univariate Marginal Distribution Algorithm (UMDA)
- The experimental results will be available soon...
Global Optimizers
Controlled Random Search (CRS)
- The experimental results will be available soon...
DIviding RECTangles Locally-biased (DIRECT-L)
- The experimental results will be available soon...
DIviding RECTangles Locally-biased RANDndomized (DIRECT-L-RAND)
- The experimental results will be available soon...
Evolutionary Strategy-Carlos Henrique (ESCH)
- The experimental results will be available soon...
Improved Stochastic Ranking Evolution Strategy (ISRES)
- The experimental results will be available soon...
Note(s):
- For these results, was used the Quantum Support Vector Classifier (QSVC) from Qiskit library, namely through the Qiskit Machine Learning module. For more information, see the following link:
Footnote(s):
License
Categories:
experimental-results
Tags:
artificial-intelligence, computer-science, quantum-support-vector-machine, support-vector-machine, quantum-kernel-training, kernel-training, quantum-kernel-alignment, kernel-alignment, quantum-machine-learning, machine-learning, supervised-learning, training, classification, iris-dataset, and intermediate