Artificial Intelligence is one such thing which is being on everybody’s lips currently, not a day goes that the AI based invention doesn’t takes place. However, the perception of people about AI is very uncertain. Some are in favour of the fact that AI is taking place of the human efforts and one day could lead to a disaster while others think AI is good for the world. What exactly is AI? There can be no such perfect definition of Artificial Intelligence or machine learning, you can say that a computer that can make a perfect chess move when the house is under pop!
A very new invention based on AI has taken place by the University of Utah researchers in United States came up with a model that can put AI into any device without any tech calculations.
New AI based model revealed
The new model is based on Artificial Intelligence that lets users to integrate artificial intelligence into any household appliance that use embedded technologies to interact with each other or with the external environment (the “Internet of Things”) which will improve data security and energy efficiency.
One more in a go
Also, the researchers of MIT (Massachusetts Institute of Technology) developed a system that came up with advanced and much smaller deep learning neural networks which can be embedded in a place like tiny computer chips in wearable medical devices, household appliances and many other object based on Internet of Things (IoT).
The system is named as the MCUNet, designed out of neural networks that deliver accuracy and remarkable speed even with limited memory and processing power.
With saving energy and data security, the technology can facilitate the expansion of the IoT universe.
MCUNet is majorly presented for deep learning and is based on two components- one is TinyEngine, an inference engine that directs resource management and is optimised to run a particular neural network structure, which is selected by the other component- TinyNAS, a neural architecture search method that creates custom-sized networks.
Then, ultimately the blend of two presents the final and efficient model to the microcontroller which will be presented at the next month’s Conference on Neural Information Processing Systems.
Use ai based kitchen to make your house more secure
MCUNet is 70.7 percent of the novel images as tested on commercial microcontroller while the previous combo of state-of-the-art neural network and inference engine was just 54 percent accurate, which indeed is a huge jump.
No more transferring data to the cloud, all your data is in safe hands now.