Nur Yahaya (fictitious person) just finished his personal shift. While he is driving constantly 12 hours a day for one of the biggest platforms, he almost earns the bare minimum to feed his family. The Malaysian depends on a system which is only possible since the upcoming smartphones. A quite easy and straight forward approach to register to the platform and he is ready to go and earn money. Hard work, pay less. Additional incomes could be charging scooters, delivering food, or walk the pets around the block.
However, there are still far more business cases for micro incomes. One of those is data labeling in computer vision. Data labeling means in terms of computer vision to describe what is being seen in a picture. Those descriptions are needed to feed machine learning models to learn to see.
For example, an autonomous driving model needs to have tons of pictures that are precisely described. The more data it gets, the more precise predictions it can make on unseen pictures/situations. These descriptions could be done on smartphones but the platform is not existing.
If we open up those people the option to earn money while waiting for the next customer, we could increase their income per hour and raise their living standard.
Some companies are trying to implement that approach, but still, they are not widely adopted today. As soon as the idea of global labeling is adopted, the machine learning industry will make a huge leapfrog. Not only the price can be lowered through lower overhead costs but also several biases can be erased by distributing the data more widely.
For the upper named reasons, we decided to start a project called Manthano. We created the first -easy to use- mobile platform, which can be used by anyone from anywhere as long they have an internet connection. It will build the foundation for the micro job platform.
