Our evidential network seamlessly transitions between learned and physics-based predictions for out-of-distribution inputs. Additionally, the physics-informed loss regularizes the learned model, ensuring better alignment with the physics model. Extensive simulations and hardware experiments demonstrate that ...
Our experiment shows that the physics loss function improves SIC and SIV predictions for most of the Arctic Ocean and winter seasons. The enhancement by the physics loss function appears more substantial when we predict SIV with a small number of training samples. 展开 ...
Multi-task learning mixture density network for interval estimation of the remaining useful life of rolling element bearings ? 2024 Elsevier LtdExisting remaining useful life (RUL) predictions of rolling element bearings have the following shortcomings. 1) Model-driven methods ty... X Wang,Y Li,K...
Furthermore, we employ our predictions to estimate, for the first time, FASER's sensitivity to electrophilic ALPs, which are predominantly generated in beauty hadron decays.Buonocore, LucaKling, FelixRottoli, LucaSominka, JonasEuropean Physical Journal C -- Parti...
The Taylor polynomial can be used to generate an approximation of the output for any input, without the use of the corresponding true labels. Therefore, the physics-based loss can be calculated for neural network predictions for all inputs. This would lead to obtaining predictions that show ...
into mind. They think that the abstract concept such as the gravitational force between apple and the earth, the curvature of earth and the fabric of space-time or black holes __ none of which has obvious application in the real world. Questions like why spend so much for looking at ...
forecasts at 30, 60, 90, and 120 minutes. It is important to note that data of NASA were not used for training. For other comparison models that cannot directly produce half-hour forecasts, we use the frame interpolation models (i.e.,FlavrandUPR) to generate 30-minute predictions. ...
The 229Th nucleus has an isomeric state at an energy of about 8 eV above the ground state, several orders of magnitude lower than typical nuclear excitation energies. This has inspired the development of a field of low-energy nuclear physics in which
There is commonly lack of uncertainty quantification for image-based solar power predictions, which however is critical to risk management in renewable-heavy grids. Information on future sky conditions, especially cloud coverage, is beneficial for PV output forecasting. With the recent advances in gener...
This method visualizes the regions of input that are “important” for predictions. The results for different samples are plotted in Fig. 4e. For small-size samples, the attention is spread over the high-order lobes. For the big-size samples, the attention is concentrated on the second-...