This page organizes papers related to artificial intelligence published around the world. This page is summarized using Google Gemini and is operated on a non-profit basis. The copyright of the paper belongs to the author and the relevant institution. When sharing, simply cite the source.
Mamba Integrated with Physics Principles Masters Long-term Chaotic System Forecasting
Created by
Haebom
Author
Chang Liu, Bohao Zhao, Jingtao Ding, Huandong Wang, Yong Li
Outline
PhyxMamba is a framework that integrates Mamba-based state-space models with physical principles to predict the long-term behavior of chaotic systems. Based on short-term state evolution observations, it reconstructs the attractor manifold using time-delayed embedding and trains Mamba to reproduce physical processes. Multi-patch prediction and attractor geometric regularization improve prediction accuracy and preserve key statistical properties of the system.
Takeaways, Limitations
•
Improving long-term prediction performance in chaotic systems by integrating Mamba-based models with physical principles.
•
Accurate predictions are possible with just short-term observation data.
•
Demonstrated excellent predictive performance in simulations and real chaotic systems.
•
The specific Limitations of the model should be further explored in the paper.
•
The complexity of the Mamba model requires consideration of computational cost and training time.
