Matching pictures and words in this platform, people can learn words in other language they want. Making labeled data by these actions, AI Company pay cost to get labeld data.

The problem Picademy solves

It is trend for many foreigner to learn Korean because of K-Culture like K-POP. And there must be some demand for learning about other language. At first in learning a language, You must know the words used in the language. These words like "고양이" as cat, "개" as dog is very easy. Probably it can be boring for you and already give up for learning new one. It must be problem.

And, in AI World, there are same problem that make you boring. To make machine learn data, You should prepare resource that will be inserted in data. Especially, AI handling picture or drawing needs labeled data which worked by human because those data should be well-refined. But the works are simple and very repeted.

Normally, you already are getting sick of works noticed above. But What would do if you can learn new words while earning some profit? We are proposing the continuous ecosystem, this L2E, which can be run by their community themselves. And this platform give a new motivation to learn new words by rewards. And we can provide labeled data made by users to AI Companay which needs labeled data.

Challenges we ran into

We are going to propose a distirubted labeled data service. And in this structure, we should fair ecosystem to our stakeholders. Especially, AI Company needs datas labeled appropriately. For running this system continuously, the both side of stakeholders should have responsibility in each their works.

In point of increasing labors for labeling data, this platform use the demand for learning new words. The users who want to know new words and at the same time and get rewards from their learning should be honorable to labeling works. And this public economy where anyone can join should have solution to handle abusing. It is important for users to give recognition they are fair.

So we implement process to check labeled data made by users. At first, we set a new role as "mentor" to check this labeled data. It seems like a 'validator' in general blockchain ecosystem. To get this role, users should stake their ERC-20 token mintied by platform in our contract. Staked ERC-20 token be used to monitoring mentor's behaviors and if the mistakes and abusing observed in this ecosystem, the appropriate amount of Staked tokens can be slashed by system.

we are implenting this level of regulation to users, but for more safe ecosystem we should develope this regulation elaborately. We are arguing that if the number of 'mentor' who check the each labeld data is increasing and divide the rewards to each 'mentor', it can be fine to continue this system.