Speaker
Yusuke Namekawa
(Kyoto Univ)
Description
We present our attempts to control the sign problem by the path optimization method with emphasis on efficiency of the neural network. We found a gauge invariant neural network is successful in the 2-dimensional U(1) gauge theory with a complex coupling. We also investigate possibility of the improvement in the learning process.
Primary author
Yusuke Namekawa
(Kyoto Univ)
Co-authors
Prof.
Kouji Kashiwa
(Fukuoka Institute of Technology)
Dr
Hidefumi Matsuda
(Fudan university)
Prof.
Akira Ohnishi
(Kyoto University)
Hayato Takase
