Optimization predicts neutrino flavor evolution, a junior prom date, and the best means to escape from an awkward party

Event Date:
2021-05-06T16:00:00
2021-05-06T17:00:00
Event Location:
Connect via zoom
Speaker:
Eve Armstrong (NYIT)
Related Upcoming Events:
Intended Audience:
Undergraduate
Local Contact:

Douglas Scott

Event Information:

The multi-messenger astrophysics of compact objects presents a vast range of environments where neutrino flavor transformation may occur and may be important for nucleosynthesis and a detected neutrino signal.  Developing efficient techniques for surveying flavor evolution solution spaces in these environments, which augment existing computational tools, could leverage progress in this field. To this end, we explore statistical data assimilation (SDA) to identify solutions to a small-scale model of neutrino flavor transformation. SDA is an optimization formula, akin to machine learning, wherein a dynamical model is assumed to generate any measured quantities.  Specifically, we use an optimization formulation of SDA wherein a cost function is extremized via the variational method. Regions of state space in which the extremization identifies the global minimum of the cost function will correspond to parameter regimes in which a model solution can exist. Our study seeks to infer the flavor transformation histories of two mono-energetic neutrino beams coherently interacting with each other and with a matter background.  We show how the procedure efficiently identifies solution regimes and rules out regimes where solutions are infeasible. Overall, results intimate the promise of this “variational annealing” methodology to efficiently probe an array of fundamental questions that traditional numerical simulation codes render difficult to access.  Finally, on a personal note, optimization has also predicted for me how things might have turned out had I mustered the nerve to ask Barry Cottonfield to the Junior Prom back in 1997, *and* it helped me sneak away from a really awkward departmental holiday party without getting caught.

Add to Calendar 2021-05-06T16:00:00 2021-05-06T17:00:00 Optimization predicts neutrino flavor evolution, a junior prom date, and the best means to escape from an awkward party Event Information: The multi-messenger astrophysics of compact objects presents a vast range of environments where neutrino flavor transformation may occur and may be important for nucleosynthesis and a detected neutrino signal.  Developing efficient techniques for surveying flavor evolution solution spaces in these environments, which augment existing computational tools, could leverage progress in this field. To this end, we explore statistical data assimilation (SDA) to identify solutions to a small-scale model of neutrino flavor transformation. SDA is an optimization formula, akin to machine learning, wherein a dynamical model is assumed to generate any measured quantities.  Specifically, we use an optimization formulation of SDA wherein a cost function is extremized via the variational method. Regions of state space in which the extremization identifies the global minimum of the cost function will correspond to parameter regimes in which a model solution can exist. Our study seeks to infer the flavor transformation histories of two mono-energetic neutrino beams coherently interacting with each other and with a matter background.  We show how the procedure efficiently identifies solution regimes and rules out regimes where solutions are infeasible. Overall, results intimate the promise of this “variational annealing” methodology to efficiently probe an array of fundamental questions that traditional numerical simulation codes render difficult to access.  Finally, on a personal note, optimization has also predicted for me how things might have turned out had I mustered the nerve to ask Barry Cottonfield to the Junior Prom back in 1997, *and* it helped me sneak away from a really awkward departmental holiday party without getting caught. Event Location: Connect via zoom