Centers & Programs

Seminars

Home Centers & Programs AI and Natural Sciences Seminars

FIELD
AI and Natural Sciences
DATE
Jan 18 (Wed), 2023
TIME
14:00 ~ 16:00
PLACE
7323
SPEAKER
Kim , Joonpyo
HOST
Lee, Jongmin
INSTITUTE
Department of Mathematics and Statistics, Sejong University
TITLE
[AI] Statistical Methodologies based on Asymmetric Losses
ABSTRACT
In the contemporary era of data science, methods based on asymmetric losses are frequently used in vast areas including statistics, deep learning, and reinforcement learning. This talk introduces several statistical methods based on the asymmetric losses with applications in several areas. Numerous existing methods rely on symmetric measures or assume symmetric structures, which fail to capture rich information on distributional characteristics lying beyond the mean or center of the data objects. To remedy this problem, some descriptive statistics and statistical approaches based on asymmetric losses, such as quantile and its variants, have been developed recently. This talk presents several asymmetric-loss-based methodologies for inhomogeneous time series or spatio-temporal data. Also, as an application, inference for tail events based on extreme value theory will also be introduced.
FILE