In the Natural Sciences, physical models are often posited and the validity of the model is assessed by comparing model predictions to experimental realizations. Such forward modeling has had its role to play and is heavily showcased throughout Physics, where disparate observations were unified into predictive frameworks inspired by logic, symmetries and other fundamental considerations. Undoubtedly, the forward approach has been tremendously successful. To wit, among many examples, this forward approach predicted the magnetic moment of the electron to a spectacular number of significant digits. But there are limitations to this historically successful approach. While forward modeling historically came first, inverse methods – spurred in equal parts by advances in probability theory (dating back to Laplace) and data-centric questions – have become indispensable as we begin to unravel the complex rules of life from single photon arrivals in Biophysics. Despite their labeling as "inverse", there is nothing backward about inferring models directly from the data. In this talk, we discuss a brief history of inverse methods in Physics and place our recent attempts at unraveling subcellular life from this perspective in our journey from Laplace's principle of indifference to Dirichlet processes.
Add to Calendar
2022-01-20T16:00:002022-01-20T17:00:00A Subjective History of Physics: from Laplace to DirichletEvent Information:
In the Natural Sciences, physical models are often posited and the validity of the model is assessed by comparing model predictions to experimental realizations. Such forward modeling has had its role to play and is heavily showcased throughout Physics, where disparate observations were unified into predictive frameworks inspired by logic, symmetries and other fundamental considerations. Undoubtedly, the forward approach has been tremendously successful. To wit, among many examples, this forward approach predicted the magnetic moment of the electron to a spectacular number of significant digits. But there are limitations to this historically successful approach. While forward modeling historically came first, inverse methods – spurred in equal parts by advances in probability theory (dating back to Laplace) and data-centric questions – have become indispensable as we begin to unravel the complex rules of life from single photon arrivals in Biophysics. Despite their labeling as "inverse", there is nothing backward about inferring models directly from the data. In this talk, we discuss a brief history of inverse methods in Physics and place our recent attempts at unraveling subcellular life from this perspective in our journey from Laplace's principle of indifference to Dirichlet processes.Event Location:
Connect via zoom