Seminar on Qualitative Theory of Differential Equations - Katarína Boďová (16.10.2025)

Thursday 16.10.2025 at 14:00, Lecture room M 223

---

Katarína Boďová:
Estimation of stochastic dynamics from tracked data

Abstract:
Advances in experimental biology now provide exceptionally rich datasets, offering new opportunities to uncover the complex dynamics driving biological systems. Yet, identifying how the underlying components interact remains a challenging inverse problem - one that requires separating systematic dynamical behavior from intrinsic stochasticity and measurement noise.

While many inverse methods exist, they often share a key drawback: they demand large, high-resolution datasets. The approach introduced by Brückner et al. (2020) offers a way to relax the first requirement. However, applying this method to settings such as population genetics - where data are typically sparse - calls for further development.

In this talk, I will present our ongoing work to refine and extend this method, addressing gaps in its original reasoning and identifying the correct small parameters that govern its validity. Our rigorous analysis not only clarifies the conditions under which the method applies but also improves its efficiency and removes the need for ad hoc assumptions - making it suitable for sparsely sampled data and a broader range of applications. This is a joint work with Richard Kollár.

 

 

 

 

---