- Developer: Milad Saadat, Deepak Mangal, Safa Jamali
- Release Date: April 27, 2024
- Model Version: 1.0
- License: MIT
- Model Type: TensorFlow implementation
- Publication: Submitted to Nature Machine Intelligence
- Primary Use: To solve integer-order and fractional fractional integro-differential equations (FIDEs) in forward and inverse directions.
- Intended Users: Researchers and professionals in fields such as physics, engineering, and applied mathematics who need to solve complex differential equations.
- Out-of-Scope Uses: The tool is not intended for non-scientific computations or for solving equations beyond the specified types of integro-differential equations.
- Inputs: Mathematical expressions of integro-differential equations, including but not limited to Fredholm and Volterra types.
- Outputs: Solutions to the input equations, presented as visual plots and data outputs within a Jupyter Notebook environment.
- Metrics: Accuracy of the solutions compared to known solutions (where available) and computational efficiency.
- Results: Demonstrated to accurately solve a variety of integro-differential equations efficiently, including test cases of integer and fractional orders.
- Care should be taken when using this model in critical applications as incorrect setup or bugs may lead to incorrect solutions that could influence subsequent decision-making.
- Users should ensure they understand the mathematical and computational complexities involved in their specific use case.
- Ensure that the TensorFlow and other dependency versions are compatible with the users’ systems to avoid computational discrepancies.
- Milad Saadat: Google Scholar Profile
- Deepak Mangal: Google Scholar Profile
- Safa Jamali: Google Scholar Profile
- Acknowledgements: Support from the National Science Foundation’s DMREF program through Award #2118962.