Artificial Intelligence is (becoming) omnipresent in our everyday lives.
Object recognition in images, natural language processing, speech recognition…AI is in our pockets, homes and vehicles.
It feels like having super-powers! (AI super-powers)
Let’s take a very basic example: I can speak a magic word and the AC units at home begin working to achieve the perfect temperature (for me).
Super-powers aside, how about having an AI-powered guardian angel at your fingertips?
How awesome would it be if I only had to pick-up my phone and look at it for a few seconds every day? The guardian angel AI can do the hard work, monitoring my fatigue and well-being. If anything is less than perfect, it provides actionable solutions, keeping me in the green and preventing me from ever having to go through a burnout. (Full disclosure: this is what we’re building at Okaya)
Where the magic happens: Cloud AI vs Edge AI
Machine learning, neural networks and deep learning are the first steps towards towards an AI-first world. The infrastructure is already in-place and constantly growing: Just think in terms of compute power, hardware efficiency and the exponential increase in data output.
AI algorithms require extensive data and processing power for best performance. So, it makes sense that most AI-enabled products and services are primarily cloud-based. The cloud has many advantages, such as simplified management and scalable computing assets, but there are some circumstances where it is not entirely adequate for AI:
- The application requires super fast responses
- The connection to the cloud service is not optimal or cannot be trusted
- The volume of data is large
The workflow of sending data to the cloud, having the data processed, and getting a response doesn’t usually take long. But pretend you are running an autopilot system; would you trust it to run effectively from the cloud?
Edge AI is not a replacement for cloud. It is a complementary approach for key business areas where using a cloud-only approach is simply not cutting it. In most of those situations a hybrid approach is best suited, where a part of the AI functions can run locally on the user’s device and the other part run in the cloud.
This is the approach we take with Okaya and by combining Edge AI with Cloud AI we aim to provide the best quality of service and user experience.