Scientists create energy-efficient AI chips inspired by the human brain
科學家研發出仿生人類大腦的高能效人工智慧晶片
In the quest for smarter technology, scientists are turning to the human brain for inspiration.
為了追求更智慧的科技，科學家們正轉向人腦尋求靈感。
Modern artificial intelligence currently relies on traditional, energy-hungry computer architectures where memory and processors are kept separate.
現代人工智慧目前依賴傳統且耗能的電腦架構，其中記憶體與處理器是分開的。
This creates a bottleneck, as constantly moving data back and forth consumes massive amounts of power.
這造成了瓶頸（ㄐㄧㄥˇ），因為不斷地來回傳輸資料會消耗龐大的電力。
Enter neuromorphic computing, a field that mimics the structure of our biological neural networks.
神經型態運算（neuromorphic computing）應運而生，這是一個模仿我們生物神經網路結構的領域。
Utilizing breakthroughs like memristors, these chips are designed to be event-driven, activating only when necessary, much like neurons firing in response to signals.
利用憶阻器（memristors）等突破性技術，這些晶片被設計為事件驅動（event-driven），僅在必要時啟動，就像神經元因應訊號而觸發一樣。
These chips are particularly promising for 'Edge AI,' powering devices like wearables and autonomous drones that require high performance on limited battery life.
這些晶片對於「邊緣AI（Edge AI）」特別有前景，能為穿戴式裝置與自動無人機等需要有限電池壽命下維持高效能的裝置提供動力。
As our need for generative AI grows, the transition to neuromorphic hardware represents a crucial step toward a more sustainable and efficient digital future.
隨著我們對生成式AI的需求增長，向神經型態硬體轉型代表著邁向更永續且高效之數位未來的重要一步。
By emulating the brain’s elegant efficiency, we are paving the way for the next generation of intelligent, low-power technology.
透過模擬大腦優雅的效率，我們正為下一代智慧、低功耗科技鋪平道路。
