From: A DeepONet multi-fidelity approach for residual learning in reduced order modeling
 |  | POD | DeepONet | MFDeepONet | ||||
---|---|---|---|---|---|---|---|---|
 |  | 10k | 20k | 50k | 10k | 20k | 50k | |
Testcase #1 | POD rank = 0.99 | 0.324 | 0.270 | 0.265 | 0.217 | 0.293 | 0.247 | 0.308 |
POD rank = 0.999 | 0.203 | 0.270 | 0.265 | 0.217 | 0.196 | 0.193 | 0.127 | |
POD rank = 0.9999 | 0.098 | 0.270 | 0.265 | 0.217 | 0.093 | 0.104 | 0.178 | |
Testcase #2 | POD rank = 0.99 | 0.105 | 0.116 | 0.068 | 0.072 | 0.030 | 0.022 | 0.030 |
POD rank = 0.999 | 0.033 | 0.116 | 0.068 | 0.072 | 0.025 | 0.023 | 0.009 | |
POD rank = 0.9999 | 0.011 | 0.116 | 0.068 | 0.072 | 0.011 | 0.009 | 0.009 |
 | Gappy POD | DeepONet | MFDeepONet | |||||
---|---|---|---|---|---|---|---|---|
 | 10k | 20k | 50k | 10k | 20k | 50k | ||
Testcase #1 | 0.260 | 0.419 | 0.380 | 0.278 | 0.218 | 0.217 | 0.197 | |
Testcase #2 | 0.135 | 0.035 | 0.033 | 0.017 | 0.029 | 0.010 | 0.009 |