세종대학교 두드림 비교과통합관리시스템

‧강연 제목 : Workshop for Korea-UK AI/ML Research in Fundamental Sciences

‧운영기간 : 2022년 10월 31일(월) 09:00 ~11월 4일 17:00

‧연사 : KIST 김경호 박사님

‧연사 일시 : 2022년 11월 2일 16:30~

‧진행 장소 : 대양 AI 빌딩 12층 컨벤션 홀 

‧신청 가능 대상 : 세종대학교 물리천문학과 학부생, 대학원생

‧신청기간 : 2022년 10월 31일(월) 09:00 ~ 11월 3일(화) 23:59까지

‧상세 내용은 물리천문학과 홈페이지를 참고하세요. 

초록

The automatic particle reconstruction with machine learning at e+e- collider We introduce the automatical particle reconstruction at e+e- collider using machine learning algorithm. This method is developed for the research based on e+e- collider. The analyses with e+e- collider are to accurately measure the Standard Model. Also, studies to find the evidence of the New Physics beyond the Standard Model are in progress. Since the characteristic of e+e- collider that a pair of B meson can be easily generated with specific beam energy, the analyses with e+e- collider are generally focused on B meson. One of the B meson is regarded and reconstructed as signal, while the other B is not reconstructed. The ‘automatic particle reconstruction’ is to reconstruct particles in the event automatically with machine learning, to build ‘other’ B meson. Since this method allows access of additional informations calculated from ‘other’ B, it is possible to improve the quality of events, though in a situation that complete reconstruction of signal B is forbidden due to invisible particles are included in the decay mode. To take the advantage of automatic particle reconstruction, a ‘tagging’ method has been used in B meson analyses. We introduce the tagging algorithm of B meson using machine learning method. Also we have applied the method of the automatic particle reconstruction, and tested the effect with one of lepton-flavor-violating decay modes with simulated samples.