Complexity-Informed Machine Learning (CIML)
PIML (physics-informed machine learning) is now being widely accepted, that is, ML indeed should respect physics. But should ML also respect chemistry, physiology, psychology ... all sciences? In particular, the reason we need ML is because the problems to solve are complex. We already advocated the triangle of "Complexity-Inverse Power Law-Fractional Calculus (C-IPL-FC)" thus it is natural to expect ML should also respect the complexity nature of the problems to solve. This page is to advocate and document Complexity-Informed Machine Learning (CIML) related research. [see also 2021 CCNC plenary talk slides ppt]
News
- 3/17/25. Mohammad Partohaghighi, Roummel Marcia, "Tail-Index-Awareness in Fractional Order Stochastic Gradient Descent" (submitted to FDTA25 under ASME/IEEE MESA 2025, part of the ASME IDETC/CIE 2025
- CIML webpage created 3/18/25
People
- Shiang Cao: Ph.D. topic - Federated Digital Twins with Complexity Informed Machine Learning (DT, CIML)
- Mohammad Partohaghighi: Ph.D. topic - More Optimal Federated Learning and Fractional Calculus (FC4FL)
- Osama Fuad Abdel Aal: Ph.D. topic - (Fractional Order) Control Theory for Machine Learning (CT4ML)
Papers/Books
- Niu, H., Chen, Y., West, B. J. (2021). Why Do Big Data and Machine Learning Entail the Fractional Dynamics?. Entropy. DOI: 10.3390/e23030297
- Haoyu Niu anf YangQuan Chen. 2024 "Smart Big Data in Digital Agriculture Applications: Acquisition, Advanced Analytics, and Plant Physiology-informed Artificial Intelligence" Springer series on Agriculture Automation and Control DOI https://doi.org/10.1007/978-3-031-52645-9
Links
- FOT4STEM book series of CRC Press. (Fractional Calculus for Skeptics I, II ...)
Created 3/18/2025 by Prof. YangQuan Chen
