'Chameleon' chip adapts to changing data speeds, cutting prediction errors by up to 40-fold

Дата публикации: 07-08-2026 13:20:04

AI semiconductors are becoming more programmable. KAIST researchers have developed a device whose response characteristics can be programmed to process data that changes at different speeds. The technology reduced prediction errors for time-varying data by up to 40-fold and is expected to enhance real-time AI performance in autonomous vehicles, robots and wearable devices.

Основное содержимое страницы с новостью.

Схожие новости

#	Наименование новости	Тональность	Информативность	Дата публикации
1	New programmable photonic chip can control how fast light moves	0	7.59	22-07-2026
2	'Oxygen tunnel' structure could solve 3D memory reliability problem to boost AI chip performance	0	9.49	19-08-2026
3	A novel chip-on-wafer platform for next-generation AI hardware	0	14.86	12-08-2026
4	Tiny memristor chip cuts brain modeling time to under 10 milliseconds	0	10.7	20-07-2026
5	'Forgetful' AI semiconductor processes recent inputs without separate resets	0	8.75	12-08-2026
6	Liquid cooling technology for semiconductor chips is 10 times more efficient than previous record	5	8	15-06-2026
7	This new brain-like chip could slash AI energy use by 70%	0	13.12	23-04-2026
8	Reconfigurable AI device preserves long-term memory while acoustic waves reset short-term signals	0	11.14	28-07-2026
9	A hardware-software co-design can efficiently run AI on edge devices	5	7	11-04-2026
10	Podcast: New microchips could supercharge AI	0	12.74	27-07-2026

Классификация: Наука. Схожих патентов: 0. Схожих новостей: 10. Тональность: 0. Информативность: 8.62. Источник: phys.org.