ASA Connect

A young researcher in continuous optimization (see https://scholar.google.cz/citations?user=Dkp_oGsAAAAJ&hl=en Click or tap if you trust this link." href="https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fscholar.google.cz%2Fcitations%3Fuser%3DDkp_oGsAAAAJ%26hl%3Den&data=02%7C01%7Cekomaroff%40keiseruniversity.edu%7Cc65e999c1bc240dc696c08d669bdfe87%7C0c38b3fe18e245159ea0b98d07b93f33%7C0%7C1%7C636812662083366861&sdata=owJ2tThU2CYZlXQjirN5ziosRHA2dLiuAfwkxLTuGqM%3D&reserved=0" target="_blank" rel="noopener noreferrer" data-auth="Verified">https://scholar.google.cz/citations?user=Dkp_oGsAAAAJ&hl=en ) has expressed interest in problems arising from state of the art novel methods in statistics. Numerical optimization is typically the last step in any procedure in statistical analysis (e.g., solving for the parameter set that minimizes a least squares function), and as in other fields (such as engineering, and the near discipline of machine learning) advances in the state of the art result in problems with new mathematical structure requiring advances in optimization algorithms, or at least careful choice and tuning thereof. For instance, if typically, for some class of new statistical methods, current solvers are not completely reliable (not always converging / without guarantees), or not as fast as one would typically like, or not even workable in many cases (i.e., due to not being able to store all the necessary information in memory for large dataset problems), this would be a fertile source of new advancement in the optimization field. Please reach out to slava@alumni.duke.edu with any open problems and challenges in this regard for pure research or applied practical collaboration.