Model the Nonlinear Universe with Machine Learning

主讲人 (Speaker): 

Yin Li

主讲人单位 (Speaker's Institute): 

东京大学数物连携宇宙研究机构

邀请人 (Invited by): 

王鑫

时间 (Time): 

星期四, 2019/06/06 - 15:00 to 16:00

地点 (Location): 

珠海校区海滨红楼17栋107 (Rm 107, Red House 17)

摘要 (Abstract): 

       Cosmological structure formation is a highly nonlinear process that forms the tiny fluctuations of the early universe into the cosmic web at the late time. Thus its accurate and efficient modeling is necessary to recover the physical information contained in the early-universe fluctuations from the late-time observables. The two conventional approaches are the numerical simulation and the perturbation theory, with the former being accurate but computationally costly, while the latter being fast but invalid below the nonlinear scale. Machine learning offers a promising third route, given its many huge successes at building nonlinear mappings. Trained with N-body simulations, our deep learning models can predict structure formation with a comparable accuracy, much higher than that of the perturbation theories at only a moderate cost. I will also discuss ongoing works that will enable a complete forward model of the observable Universe.

主讲人简介 (Speaker's CV): 

Yin Li 在芝加哥大学获得博士学位，然后作为一名联合博士后，在加州大学伯克利分校和东京大学数物连携宇宙研究机构工作。Yin Li目前在东京大学数物连携宇宙研究机构工作，从今年秋天开始，将以研究员的身份在纽约工作。