[15a-A205-1~10] 23.1 Joint Session N "Informatics"

▲ [15a-A205-7] Search of magnetocaloric materials through machine-learning

〇Pedro Castro^1,2, Kensei Terashima¹, Takafumi D Yamamoto¹, Suguru Iwasaki¹, Ryo Matsumoto^1,2, Shintaro Adachi¹, Saito Yoshito^1,2, Song Peng^1,2, Hiroyuki Takeya¹, Yoshihiko Takano^1,2 (1.MANA, NIMS, 2.Tsukuba Univ.)

Keywords:magnetocaloric effect, machine learning

Magnetocaloric effect is emerging as a potential substitute for conventional gas-cycle based refrigerators, due to its enviromental friendliness and effectiveness, being suitable for cooling from room temperature to liquefaction of gases. However, for exploiting such effect, materials that exhibits remarkable magnetic entropy change are needed.
In this talk, we show our current efforts into modeling the magnetocaloric effect of magnetic materials through use of machine learning, discussing also ongoing experimental validation of such method and the challenges of this data-driven approach.