Deep learning models can be improved for physical processes by incorporating prior scientific knowledge. The paper proposes a method where a neural network predicts parameters like motion fields in governing equations, rather than directly predicting outputs. It applies this to ocean surface temperature prediction. The model predicts a motion field from past temperature images using a CNN. It then uses this motion field in a warping scheme based on the advection-diffusion equation to forecast future temperature. This outperforms comparison methods by leveraging physics knowledge without requiring manual specification of equations.
Muon g-2 anomaly
- The measured value of the muon anomalous magnetic moment (g-2) differs from the standard model prediction by more than 3 sigma, indicating potential new physics.
- Future experiments at Fermilab and J-PARC aim to measure the muon g-2 with higher precision and could discover a discrepancy exceeding 5 sigma in the next 3-5 years.
- This research focuses on investigating the source of the current muon g-2 anomaly by considering contributions from supersymmetric models, which could accommodate new particles within the mass reach of current accelerators.
The document discusses prospects for slepton searches in future experiments based on previous works. It summarizes the status of the muon anomalous magnetic moment measurement, which shows a 3-4 sigma discrepancy from the Standard Model prediction. This discrepancy could be explained by contributions from new physics, such as supersymmetry. Supersymmetry predicts superpartner particles like sleptons. The author's dissertation will examine the muon g-2 anomaly within the framework of the minimal supersymmetric standard model and study slepton mass bounds and prospects for discovering sleptons at future colliders in a model-independent way to help explain the muon g-2 discrepancy.
Deep learning models can be improved for physical processes by incorporating prior scientific knowledge. The paper proposes a method where a neural network predicts parameters like motion fields in governing equations, rather than directly predicting outputs. It applies this to ocean surface temperature prediction. The model predicts a motion field from past temperature images using a CNN. It then uses this motion field in a warping scheme based on the advection-diffusion equation to forecast future temperature. This outperforms comparison methods by leveraging physics knowledge without requiring manual specification of equations.
Muon g-2 anomaly
- The measured value of the muon anomalous magnetic moment (g-2) differs from the standard model prediction by more than 3 sigma, indicating potential new physics.
- Future experiments at Fermilab and J-PARC aim to measure the muon g-2 with higher precision and could discover a discrepancy exceeding 5 sigma in the next 3-5 years.
- This research focuses on investigating the source of the current muon g-2 anomaly by considering contributions from supersymmetric models, which could accommodate new particles within the mass reach of current accelerators.
The document discusses prospects for slepton searches in future experiments based on previous works. It summarizes the status of the muon anomalous magnetic moment measurement, which shows a 3-4 sigma discrepancy from the Standard Model prediction. This discrepancy could be explained by contributions from new physics, such as supersymmetry. Supersymmetry predicts superpartner particles like sleptons. The author's dissertation will examine the muon g-2 anomaly within the framework of the minimal supersymmetric standard model and study slepton mass bounds and prospects for discovering sleptons at future colliders in a model-independent way to help explain the muon g-2 discrepancy.