This page evolves as fast as our algorithms do. 🚀

We are a research group at the forefront of scientific machine learning, dedicated to building scalable, physics-inspired machine learning algorithms that accelerate simulation, prediction, and decision-making in complex physical systems. Our core philosophy: "Let physics guide the learning, and let learning scale the physics-based solvers."

CSCCM's work bridges theory and practice through innovations in physics-enhanced machine learning, computational mechanics, and high-performance computing. Our group is primarily working in three thrust areas:

• Thrust Area I: Bridging Scientific Discovery and Machine Learning: In this thrust area, we aim to bridge scientific discovery and machine learning. Some of the representative work carried out by our group in this area includes, (a) Discovering SDE from data, (b) AI for solving physical reasoning tasks, (c) AI-accelerated topology optimization, (d) Reinforcement learning for vibration control, among others.

• Thrust Area II: Scientific Machine Learning at Scale: In this thrust area, we work on developing scalable scientific machine learning algorithms. We are particularly interested in developing algorithms that can handle sparse and tall/slender data problems. Some of the representative work in this thrust area includes, (a) Wavelet Neural Operator, (b) Deep Physics Corrector, (c) Foundation model for physical systems, and (d) neurosciene-inspired neural models.

• Thrust Area III: Uncertainty Quantification And Reliability Analysis: Lastly, in the thrid thrust area, we work on UQ for ML and ML for UQ. Some representative work includes, (a) Gaussian Process Operator, (b) Distribution-free UQ in neural operator, and (c) Time-dependent reliability analysis with neural operator.

The three thrust areas closely interacts with each other. We aim for a future where machines autonomously discover physical laws, reason about the unknown, and collaborate with humans in real-time scientific exploration

For further details, please visit www.csccm.in. You can check our publications at this link

If you are interested in our work, you may also be interested in the SciML webinar series JHU-IITD SMaRT that we are hosting in collaboration with Prof. Somdatta Goswami from Johns Hopkins University and my colleague Prof. Rajdip Nayek from the Indian Institute of Technology Delhi.