AI-Based Reduced Order Modeling at MATLAB EXPO
Kishen Mahadevan
on 7 Nov 2025 at 18:41
Latest activity Reply by Mahaveer Satra
on 14 Nov 2025 at 15:13
Hi everyone!
I’m Kishen Mahadevan, Senior Product Manager at MathWorks, where I focus on controls and deep learning. I’m excited to be speaking at MATLAB EXPO this year!
In one of my sessions, I’ll share how AI-based reduced order models (ROMs) are transforming engineering workflows—using battery fast charging as an example—making it easier to reuse high-fidelity models for real-time control and deployment.
I’d love to have you join the conversation at the EXPO and right here in the community!
Feel free to drop any questions or thoughts ahead of the event.
4 Comments
Time Descending
Can the training of the AI based data driven ROM be sped up using GPU based computational resources?
Yes you can use multiple GPUs to accelerate the training process for creating the ROM. This resource walk through the workflow: https://www.mathworks.com/help/deeplearning/ug/custom-training-experiment-using-parallel-workers.html
The ROMs session looks great. Are you posting the source code to GitHub/File Exchange? Thanks.
Yes, you can access the workshop files from here: https://github.com/MahaveerSatra/Reduced-Order-Model-Hands-on-Workshop---AI-with-Model-Based-Design
There are more useful exmaples here: https://www.mathworks.com/matlabcentral/fileexchange/156364-reduced-order-modeling-support-package-beta#examples_tab
Sign in to participate
