Summary of Learning Hidden Physics and System Parameters with Deep Operator Networks, by Vijay Kag et al.
Learning Hidden Physics and System Parameters with Deep Operator Networks
by Vijay Kag, Dibakar Roy Sarkar, Birupaksha Pal, Somdatta Goswami
First submitted to arxiv on: 6 Dec 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: None
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The paper proposes two neural operator frameworks to discover hidden physics and identify unknown system parameters from sparse measurements. The first framework combines a popular neural operator, DeepONet, with a physics-informed neural network to accurately capture the relationship between sparse data and underlying physics. The second framework uses a pre-trained DeepONet on sparse sensor measurements to initialize a physics-constrained inverse model for parameter identification. Both frameworks excel in handling limited data and preserving physical consistency. The paper showcases state-of-the-art performance on benchmarking tasks, achieving average L2 errors of O(10^-2) for hidden physics discovery and absolute errors of O(10^-3) for parameter identification. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper introduces two new neural operator frameworks to help scientists understand complex systems better. It combines a popular model called DeepONet with another type of AI, physics-informed neural networks, to figure out how sparse data relates to the underlying laws of nature. The second framework uses this combination to find unknown parameters in a system from limited measurements. Both models work well even when there’s not much data and stay true to the physical rules we know. The paper shows that these frameworks can solve complex problems really well, with errors as small as 0.01 for some tasks. |
Keywords
» Artificial intelligence » Neural network
