We explore the application of convolutional neural networks (CNNs) to 21cm maps as a tool for probing dark matter properties. Following a review of 21cm physics and CNN algorithms, we discuss three concrete dark matter scenarios: ultra-light dark matter (axion-like particles), annihilating dark matter, and warm dark matter.
We demonstrate the utility of CNNs in differentiating dark matter models by linking 21cm map features to the underlying physics of dark matter. Our study highlights image-based cosmology as a novel framework for studying dark matter.
Bio:
PhD in Physics (2004), University of California Berkeley, USA
BSc in Mathematical Physics (1999), University of Manchester Institute of Science and Technology, UK
After obtaining the PhD at UC Berkeley, I worked as a postdoctoral fellow at Fermilab (theoretical astrophysics group), Minnesota (particle theory group) and Michigan (particle phenomenology group). I was the research assistant professor at Nagoya University (theoretical cosmology group) in Japan and then the faculty scientist at the Institute for Basic Science (IBS) in South Korea. I joined HIAS-UCAS/ICTP-AP in Hangzhou, China as a senior faculty scientist in 2021.
Alternative link: https://meeting.tencent.com/dm/NMFpcWhWilEl
#腾讯会议：904538437
会议密码：无
Jason Evans