STM Journals Peer-reviewed research from Consortium e-Learning Network

Research & Reviews : Journal of Physics Original Research

Harnessing Artificial Intelligence for Precision Physics: A Machine Learning Framework for Data Reconstruction in Support of India's Deep-Tech Missions

  1. Shubham Kumar Modi Radha Govind UniversityRamgarh, Jharkhand 1Supervisor, Jhumri Telaiya Municipal Council, Koderm

Abstract

India's emergence as a global leader in deep-tech innovation is driven by ambitious scientific megaprojects, including the Laser Interferometer Gravitational-Wave Observatory (LIGO)-India, the X-ray Polarimeter Satellite (XPoSat), the Aditya-L1 solar observatory, and the National Quantum Mission (NQM). However, the unprecedented scale and complexity of the observational data generated by these missions present severe computational bottlenecks. Traditional analytical frameworks struggle with non-stationary noise transients, diffusion blurring, and the exponential scaling limits inherent in multidimensional physical systems. This paper proposes a comprehensive, publicationready machine learning framework specifically engineered for precision physics data reconstruction. The tripartite framework integrates Physics-Informed Neural Networks (PINNs) and Physics- Informed Kolmogorov-Arnold Networks (PIKANs) to enforce strict adherence to governing physical laws; selective State Space Models (SSMs), specifically the Mamba architecture, to achieve lineartime sequence modeling for continuous telemetry; and Geometric Deep Learning (GDL), including Hypernetwork-modulated Restricted Boltzmann Machines (HyperRBMs), to preserve spatial symmetries in non-Euclidean domains. Our analysis demonstrates that the deployment of reinforcement learning via Deep Loop Shaping successfully reduces optomechanical mirror jitter in gravitational-wave interferometers by a factor of 30 to 100 compared to classical controllers. Furthermore, the application of deep convolutional ensembles for X-ray polarimetry reduces required observation exposure times by approximately 40%. In the quantum domain, parametric Quantum State Tomography (QST) utilizing GDL achieves over 90% reconstruction fidelity for multi-qubit systems while completely bypassing the traditional exponential measurement barrier. By transitioning artificial intelligence from a passive post-processing tool to an integrated, physics-constrained component of experimental instrumentation, this framework provides the essential computational scaffolding required to maximize the scientific yield of India's sovereign deep-tech megaprojects through 2026 and beyond.

Keywords

References (38)

  1. Chatterjee C, Jani K. Reconstruction of binary black hole harmonics in LIGO using deep learning.
  2. Astrophys J. 2024;969(1):25. doi:10.3847/1538-4357/ad4602.
  3. LIGO Laboratory. Artificial intelligence helps boost LIGO. Pasadena (CA): LIGO Caltech; 2025.
  4. Available from: LIGO News Article.
  5. Innan N, Siddiqui OI, Arora S, Ghosh T, Koçak YP, Paragas D, et al. Quantum state tomography
  6. using quantum machine learning. Quantum Mach Intell. 2023;5:10327.
  7. doi:10.48550/arXiv.2308.10327.
  8. Press Information Bureau, Government of India. National Quantum Mission: India's quantum
  9. leap. New Delhi: Press Information Bureau; 2025. Available from: PIB Release.
  10. Baldini L, Negro M, Moriakov N, et al. A deep ensemble approach to X-ray polarimetry. In:
  11. Advances in Neural Information Processing Systems. Vol. 34. Red Hook (NY): Curran Associates
  12. Inc.; 2021. p. 1-10.
  13. Cibrario N, Negro M, Moriakov N, Bonino R, Baldini L, Di Lalla N, et al. Joint machine learning
  14. and analytic track reconstruction for X-ray polarimetry with gas pixel detectors. Astron
  15. Astrophys. 2023;674:A107. doi:10.1051/0004-6361/202346302.
  16. Physics-informed neural networks: a review of methodological approaches and applications. Appl
  17. Sci. 2025;15(14):8092.
  18. Gu A, Dao T. Mamba: linear-time sequence modeling with selective state spaces [Preprint]. 2023.
  19. arXiv:2312.00752.
  20. Mamba, which is integrated with physics principles, masters long-term chaotic system
  21. forecasting. In: OpenReview Proceedings. 2025. Available from: OpenReview Paper.
  22. LongMamba: enhancing Mamba's long-context capabilities via training-free receptive-field
  23. enlargement. In: OpenReview Proceedings. 2024. Available from: OpenReview LongMamba.
  24. Equivariant graph neural networks for charged particle tracking [Preprint]. 2023.
  25. arXiv:2304.05293.
  26. Jabeen S, Kurdydyk D, Palnitkar A, Talati M, Yan J, Yang J. Quantum machine learning for state
  27. tomography using classical data. Quantum Mach Intell. 2026;8:40. doi:10.1007/s42484-026-
  28. HEASARC, NASA. XPoSat mission overview. Greenbelt (MD): NASA. Available from: XPoSat
  29. HEASARC Page.
  30. Indian Space Research Organisation. Aditya-L1 decodes the impact of powerful solar storm on
  31. Earth's invisible magnetic shield. Bengaluru: Indian Space Research Organisation; 2025.
  32. Available from: Aditya-L1 Solar Storm Report.
  33. Quantum Zeitgeist. IISc builds 6-qubit photonic quantum system—a first for India. 2025.
  34. Available from: IISc Quantum System News.
  35. Department of Science and Technology, Government of India. National Quantum Mission
  36. (NQM). New Delhi: Department of Science and Technology. Available from: DST National
  37. Quantum Mission.
  38. Aditya-L1. In: Wikipedia. Wikimedia Foundation; 2026. Available from: Aditya-L1 Wikipedia
Support