Predicting the importance of complicated regions in Monte Carlo sampling

Nov 13, 2023, 8:50 PM
20m
The Suites Hotel, Jeju Island, Korea

The Suites Hotel, Jeju Island, Korea

67 Jungmungwangwang-ro 72beon-gil, 특별자치도, Seogwipo-si, Jeju-do

Speaker

Raymundo Ramos (KIAS)

Description

In this talk we discuss a proposal where a neural network is trained to classify regions in multidimensional parameter space according to a function that determines their importance. These regions in general can have complicated shapes and disconnected subregions. Accurate classification of regions of importance with a neural network can improve event generation and Monte Carlo integration, specially for costly time consuming calculations.

Primary author

Presentation materials