George Stepaniants
Scholar

George Stepaniants

Google Scholar ID: CKYZLxYAAAAJ
Computing and Mathematical Sciences, California Institute of Technology
Machine LearningStatisticsInverse ProblemsMathematical Physics
Citations & Impact
All-time
Citations
144
 
H-index
5
 
i10-index
5
 
Publications
11
 
Co-authors
20
list available
Resume (English only)
Academic Achievements
  • Invited Speaker, Banff BIRS Workshop on Efficient and Reliable Deep Learning (June 2025).
  • Invited Speaker, Fields Institute Symposium on Machine Learning and Dynamical Systems (June 2025).
  • NSF Mathematical Sciences Postdoctoral Research Fellowship (MSPRF) awardee (September 2024).
  • Recipient of IMS Lawrence D. Brown Ph.D. Student Award (one of three nationwide, September 2024).
  • Presented at Armenian Statistics Summer School under Calouste Gulbenkian Travel Grant (June 2023).
  • Invited Speaker, CIRM Meeting on Mathematical Statistics (December 2021).
  • NSF Graduate Research Fellowship (GRFP) and MIT Presidential Fellowship recipient (September 2019).
  • Elected to Phi Beta Kappa Honors Society (June 2019).
Background
  • Develops methods to learn mathematical and physical laws from simulated and experimental data, integrating numerical analysis, mechanics, statistics, and machine learning to succeed in data-limited regimes with scientifically informed inductive biases.
  • Addresses three core challenges: (I) alignment and pooling of scientific data from disparate sources; (II) inference of governing physical laws under noisy and sparse data; (III) design of memory-dependent (autoregressive) and higher-order models to compensate for partial observability in dynamical systems.
  • Applies methodological advances to domains including biochemistry, materials science, and fluid mechanics.
  • Teaching philosophy emphasizes guiding students to discover mathematical ideas in domain-specific literature, formalize them into well-posed theories, and implement them as reproducible numerical algorithms.
  • Seeking tenure-track positions in applied and computational mathematics, data science, statistics, and engineering departments during 2025–2027.