KAIST develops AI chip that recognizes changes in motion over time

KAIST researchers developed a stacked semiconductor AI chip using solid ion-gel thin films that can recognise motion changes over multiple timescales. The chip achieves over 90% accuracy in classifying moving image sequences, operates stably on flexible substrates, and is aimed at low-power devices like smartwatches.

This Wire brief sits within Fusion42's coverage of AI & ML.