- 일시: 2024년 12월 23일 10시
- 장소: R관 715호
- 연사: UPenn CS 박사과정 김규래 연구원
- 제목: Wideband Signal Detection via Bayesian
Model Comparison
- 초록:
In array signal processing, determining the
number of incoming signals is known as the signal detection problem. Unlike the
more typical direction-of-arrival estimation problem, the signal detection problem
is more challenging as it corresponds to a model comparison problem, where we
must answer the more fundamental question of "What is a signal?"
Among various approaches to model comparison, information-theoretic methods and
hypothesis testing have been extensively studied. In contrast, the Bayesian
approach to model comparison has been relatively less explored, especially for
wideband array signal processing. In this talk, I will demonstrate that, with
proper probabilistic modeling, Bayesian model comparison can significantly
outperform alternatives in the low snapshot-low SNR regime.
- Reference:
https://arxiv.org/abs/2412.08895
- Bio:
Kyurae Kim is a third-year Ph.D. student at
the University of Pennsylvania, advised by Professor Jacob R. Gardner. His
research focuses on Bayesian inference, stochastic optimization, MCMC methods,
Bayesian optimization, and their signal processing applications. He
previously worked as a research associate at the University of Liverpool and
holds a Bachelor of Engineering from the Electronics Engineering Dept. of
Sogang University, South Korea.
- 주관: 김홍석 교수 연구실