[22p-M206-1~16] 23.1 Joint Session N "Informatics"

[22p-M206-11] Toward robust and fast neural networks learning for efficient use of repository and preliminary physical properties measurement data

〇Kensei Terashima¹, Pedro Baptista de Castro^1,2, Miren Garbine Esparza Echevarria^1,2, Ryo Matsumoto¹, Takafumi D. Yamamoto¹, Hiroyuki Takeya¹, Yoshihiko Takano^1,2 (1.NIMS, 2.Univ. of Tsukuba)

Keywords:neural networks, machine learning, Efficient measuremtn

Interpolation technology that can accurately follow unequally spaced and non-linear changes is desired as a tool for effectively utilizing the repository and preliminary data to build an efficient experimental plan. In this talk, we will report our efforts to quickly and accurately perform fully-connected feedforward neural networks learning, during an experiment on-the-fly.