Sustainable edge intelligence
Inference accelerator for the AI of tomorrow
Artificial intelligence - both training and the actual inference - could so far mainly be developed for data centres. This trend is currently changing with the new emerging field of "Edge AI". Smartphones, robots, drones, surveillance cameras and industrial cameras - all of these devices will have some kind of AI processing built into them in the near future.
When inference is to take place directly on the imaging device itself, it will get interesting. How can such a high-power technology be used efficiently and sustainably outside of large data centres in the small resource-optimised embedded devices?
In our technical article "Sustainable Edge Intelligence" we show you possible hardware options and the challenges they pose in order to move AI "on the edge". You will also get a basic understanding of how the deep ocean core CNN accelerator, which is integrated into the IDS NXT camera family, works.