|
Speaker |
Movahed, Seyed Mohammad Sadegh |
|
Coauthors |
Alireza Vafaei Sadr, Marzieh Farhang, Christophe Ringeval, François R. Bouchet, Bruce Bassett, Martin Kunz |
| Talk Title |
Multi-scale searching machine to detect the cosmic strings network |
|
Abstract |
Cosmic topological defects formed during phase transition in the very early universe are theoretically well-motivated. The cosmic string (CS) can leave the imprint on the CMB stochastic field leading to emerge additional stochasticity behavior. We will rely on the stochasticity nature of CMB superimposed by Cs network and therefore some topological and geometrical measures accompanying multi-scale edge-detection algorithm to examine the capability of Cs detection, will be introduced and utilized. On the noiseless sky maps with an angular resolution of $0.9'$, we show that our pipeline detects cosmic string with $G\mu$ as low as $G\mu> 4.3\times 10^{-10}$. We also explore two powerful tree-based machine learning algorithms to exploit the feature importance. Such approach opens new insight into utilizing prior information for detecting exotic features in a cosmological stochastic field. Machine learning method enables us to detect Cs with $G\mu>2.1\times 10^{-10}$ at $3\sigma$ level. |
|
Talk view |
|
|
|
|