Physical System Modeling on Dataset 1
0.0064RMSERNN_uncertainty
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| RNN_uncertaintyModel Family=RNN Models, PIF Strategy=dynamic uncertainty-based loss, Approx. Parameters=30,000, Time(s)=7.22025.09 | 0.0064 | 0.0333 | 0.0019 | 0.0163 | 0.4251 | 0.0454 | |
| RNN_RBAModel Family=RNN Models, PIF Strategy=Residual-Based Attention, Approx. Parameters=30,000, Time(s)=8.72025.09 | 0.0085 | 0.035 | 0.0019 | 0.0154 | 0.5313 | 0.042 | |
| cKANModel Family=cKAN Models, PIF Strategy=None, Approx. Parameters=30,000, Time(s)=9.92025.09 | 0.0086 | 0.0705 | 0.0017 | 0.0145 | 0.2125 | 0.0308 | |
| LEM_fixedModel Family=LEM Models, PIF Strategy=fixed-weight loss, Approx. Parameters=30,000, Time(s)=29684.62025.09 | 0.0088 | 0.079 | 0.0018 | 0.0153 | 0.4251 | 0.0287 | |
| KANModel Family=KAN Models, PIF Strategy=None, Approx. Parameters=30,000, Time(s)=1835.92025.09 | 0.0093 | 0.0466 | 0.002 | 0.0189 | 1.2221 | 0.0225 | |
| cPKAN_fixedModel Family=cKAN Models, PIF Strategy=fixed-weight loss, Approx. Parameters=30,000, Time(s)=109.92025.09 | 0.0094 | 0.0549 | 0.0016 | 0.0158 | 0.0531 | 0.0241 | |
| PKAN_fixedModel Family=KAN Models, PIF Strategy=fixed-weight loss, Approx. Parameters=30,000, Time(s)=29389.32025.09 | 0.0094 | 0.048 | 0.002 | 0.0189 | 1.2221 | 0.0224 | |
| RNNModel Family=RNN Models, PIF Strategy=None, Approx. Parameters=30,000, Time(s)=12.22025.09 | 0.0097 | 0.0395 | 0.0019 | 0.0164 | 0.4782 | 0.0304 | |
| LSTM_fixedModel Family=LSTM Models, PIF Strategy=fixed-weight loss, Approx. Parameters=30,000, Time(s)=185.52025.09 | 0.0097 | 0.0569 | 0.0019 | 0.0154 | 0.3188 | 0.0331 | |
| MLPModel Family=MLP Models, PIF Strategy=None, Approx. Parameters=30,000, Time(s)=10.32025.09 | 0.0097 | 0.066 | 0.0017 | 0.0181 | 0.1594 | 0.0215 | |
| RNN_fixedModel Family=RNN Models, PIF Strategy=fixed-weight loss, Approx. Parameters=30,000, Time(s)=155.82025.09 | 0.01 | 0.0378 | 0.0019 | 0.015 | 0.4782 | 0.0319 | |
| LEM_RBAModel Family=LEM Models, PIF Strategy=Residual-Based Attention, Approx. Parameters=30,000, Time(s)=2537.42025.09 | 0.0103 | 0.061 | 0.0018 | 0.0158 | 0.3719 | 0.0262 | |
| LSTMModel Family=LSTM Models, PIF Strategy=None, Approx. Parameters=30,000, Time(s)=15.62025.09 | 0.0106 | 0.0735 | 0.0019 | 0.0156 | 0.3188 | 0.034 | |
| LSTM_RBAModel Family=LSTM Models, PIF Strategy=Residual-Based Attention, Approx. Parameters=30,000, Time(s)=13.72025.09 | 0.0112 | 0.0598 | 0.002 | 0.015 | 0.4782 | 0.0385 | |
| LSTM_uncertaintyModel Family=LSTM Models, PIF Strategy=dynamic uncertainty-based loss, Approx. Parameters=30,000, Time(s)=10.02025.09 | 0.0112 | 0.0622 | 0.002 | 0.0153 | 0.3719 | 0.0354 | |
| LEM_uncertaintyModel Family=LEM Models, PIF Strategy=dynamic uncertainty-based loss, Approx. Parameters=30,000, Time(s)=1448.92025.09 | 0.0115 | 0.065 | 0.0017 | 0.0162 | 0.2657 | 0.0274 | |
| LEMModel Family=LEM Models, PIF Strategy=None, Approx. Parameters=30,000, Time(s)=1447.22025.09 | 0.0119 | 0.0864 | 0.0018 | 0.0156 | 0.4251 | 0.017 | |
| PKAN_uncertaintyModel Family=KAN Models, PIF Strategy=dynamic uncertainty-based loss, Approx. Parameters=30,000, Time(s)=1808.72025.09 | 0.0147 | 0.094 | 0.0019 | 0.0202 | 0.4782 | 0.0185 | |
| PKAN_RBAModel Family=KAN Models, PIF Strategy=Residual-Based Attention, Approx. Parameters=30,000, Time(s)=1963.82025.09 | 0.0148 | 0.1061 | 0.0021 | 0.0209 | 1.0627 | 0.0232 | |
| MLP_uncertaintyModel Family=MLP Models, PIF Strategy=dynamic uncertainty-based loss, Approx. Parameters=30,000, Time(s)=10.32025.09 | 0.0151 | 0.0834 | 0.0018 | 0.0189 | 0 | 0.0226 | |
| Transformer_fixedModel Family=Transformer Models, PIF Strategy=fixed-weight loss, Approx. Parameters=30,000, Time(s)=647.72025.09 | 0.0165 | 0.0659 | 0.0015 | 0.0258 | 0.0531 | 0.0212 | |
| MLP_fixedModel Family=MLP Models, PIF Strategy=fixed-weight loss, Approx. Parameters=30,000, Time(s)=102.42025.09 | 0.0169 | 0.121 | 0.0018 | 0.0192 | 0 | 0.0191 | |
| Transformer_RBAModel Family=Transformer Models, PIF Strategy=Residual-Based Attention, Approx. Parameters=30,000, Time(s)=66.82025.09 | 0.0172 | 0.0977 | 0.0016 | 0.0397 | 0 | 0.029 | |
| MLP_RBAModel Family=MLP Models, PIF Strategy=Residual-Based Attention, Approx. Parameters=30,000, Time(s)=8.62025.09 | 0.0177 | 0.0858 | 0.0018 | 0.0185 | 0 | 0.0295 | |
| cPKAN_RBAModel Family=cKAN Models, PIF Strategy=Residual-Based Attention, Approx. Parameters=30,000, Time(s)=9.62025.09 | 0.0199 | 0.0886 | 0.0017 | 0.0176 | 0.3719 | 0.0431 | |
| Transformer_uncertaintyModel Family=Transformer Models, PIF Strategy=dynamic uncertainty-based loss, Approx. Parameters=30,000, Time(s)=31.62025.09 | 0.0199 | 0.1205 | 0.0015 | 0.0171 | 0.1063 | 0.0303 | |
| cPKAN_uncertaintyModel Family=cKAN Models, PIF Strategy=dynamic uncertainty-based loss, Approx. Parameters=30,000, Time(s)=6.42025.09 | 0.0205 | 0.1257 | 0.0018 | 0.0153 | 0.1063 | 0.0316 | |
| TransformerModel Family=Transformer Models, PIF Strategy=None, Approx. Parameters=30,000, Time(s)=46.82025.09 | 0.0206 | 0.119 | 0.0015 | 0.0285 | 0.2125 | 0.0145 | |
| Recipe Prior BaselineModel Family=Baseline, Approx. Parameters=30,0002025.09 | 0.0259 | 0.1618 | 0.0016 | 0.021 | 0 | 0 |