Physics-based AI model opens new frontiers in dielectric materials exploration
Get the Health newsletter
Daily health & science — research, biotech, public health, the studies worth knowing. Free.
- Atsushi Takigawa leads the Tohoku University team that built the physics‑guided AI model for dielectric screening.
- Tohoku University researchers integrated AI with Born effective charge and phonon calculations, achieving higher accuracy than conventional methods.
- Physical Review X published the study, validating the factorized machine‑learning approach for predicting ionic dielectric tensors.
- Screening of >8,000 oxides identified 31 previously unknown high‑dielectric oxide materials, promising smaller, more efficient electronic components.
Why it matters: The 31 newly identified high‑dielectric oxides could enable smaller, more efficient capacitors for smartphones and computers.
