Source
Nowadays most of the tech uses machine learning models to enhance their performance or gain additional features. Even without us realizing Ml is being implemented in our life like you reading this blog because you came across it. This blog was on your dashboard due to the ML learning model which learned what type of blogs you read 馃摉.
These ml models take a heavy toll on the computer hardware so they are usually computed on extended GPU or cloud servers. When we want some process such as Google assistant on phone to work instantaneously even without internet but that's not possible since our voice-first goes to the Google server where it is analyzed and an appropriate response is sent back this spoils the user XP(experience).
To solve this problem a new field on ML was created called "Tiny ML", here basically we explore various models that can be run on smaller devices, with not very high computing capabilities such as microprocessors which take 200million times less voltage than that a GPU takes(200-500 watts).
How exactly is tiny ML applied in our day to day life
鉁旓笍Even though we don't realize we have been utilizing tiny ML in our life just like how we have been using various ML learning models. The words such as "Alexa" and "Ok Google" which are used to turn on iot baaed devices these use tiny ML during the wake-up and in other basic functions; Here they learn to recognize your voice and are always trying to capture equivalent to your voice saying "ok google".
鉁旓笍Solar Scare Mosquito 馃 project uses tiny Ml to find breeding spots of mosquitoes and then agitates water over those locations. This was if more mosquitoes do not breed the chance of being affected by mosquito-borne diseases such as dengue, zika virus, malaria, etc can be avoided. Also as the name states " Solar" is a solar energy-based project and can run indefinitely inappropriate conditions.
Source
鉁旓笍 Nowadays apps can fine diseases and defects in plants by just taking a picture of the plant and then running through their data sets present on the app to find out what kind of disease it is. This benefits the farmers hugely since they can figure out diseases at an early stage and prevent it. Here since tiny Ml is used it can be utilized even without an internet connection which is very helpful since many farmlands have very bad network.
鉁旓笍ML based devices are made in the seaports where there is a chance of whale strike. Here they can be altered on time to prevent the loss of whales.
I would like to conclude by saying that Tiny ML is used in so many places especially where microprocessors are utilized where they collect a small amount of data that could be used to develop better applications or enhance their performance. Tiny ML as a field definitely has a bright future ahead.
Hope You learned something new in this blog, Let me know your thoughts in the comments :^)
