Blockchain based Decentralized Machine learning protocols and marketplaces that reward users

in #blog4 years ago


Numerai is a platform, that rewards data scientists and assembles all models into a meta-model which is then applied to a centralized hedge fund. Based on the applicability of the submitted ML algorithm, users are rewarded. Also, users who submit their ML algorithms have to stake their corresponding cryptocurrencies in NMR onto the platform. More recently, the team behind Numerai released a protocol called Erasure, which creates a decentralized marketplace for submission and solicitation of machine learning or other forms of data. These data and algorithms marketplaces are completely decentralized having their own governance structure and rewards structure enabling peer-to-peer trade.

These decentralized machine learning (ML) marketplaces combine private machine learning and open blockchain algorithms, and enable crowdsourcing of models, without need for disclosing data (by using homo-morphic encryption). Such an arrangement is a powerful harbinger to the Artificial intelligence revolution being enabled by web 3.0 - similar to the open-source revolution that drove most of the infrastructure used in the internet industry - e.g., Linux, Apache, WordPress, etc...

The value deriving mechanism in such a decentralized machine learning marketplace moves from data to the algorithms and the efficacy of the models, which can now be crowd-sourced without having to reveal data. An added bonus is the ability to reward algorithm creators based on the resulting efficiency of the model in the real world operation. For example, on the Numerai platform, many individuals would have submitted algorithms with corresponding stakes, however, only those algorithms which perform (or return the highest) the best are rewarded with a percentage of the profits proportional to their model's contribution. The bad algorithms are not rewarded, and very often the submitter loses their stakes through a process of burning on the blockchain. In the web 2.0 era, firms (large and small) possessed extremely sensitive and often private non-replicable data that was used to increase their competitive advantage. As a result, these firms focused on hiring the best of the talent from the markets to work for them.

With the evolving technology trends of Numerai, homo-morphic encryption, and the erasure protocol now – in addition to data – algorithms, models, and all kinds of services can be monetized directly in a peer-to-peer fashion. One of the early marketplaces using the NMR token from Numerai and rewarding both data scientists and algorithm creators, in addition to all kinds of data and algorithmic work is this website called erasurebay.org.



Posted from my blog with SteemPress : https://www.cryptonewtech.com/2020/08/09/blockchain-based-decentralized-machine-learning-protocols/
Sort:  

Congratulations @enterprof! You have completed the following achievement on the Hive blockchain and have been rewarded with new badge(s) :

You published more than 80 posts. Your next target is to reach 90 posts.
You received more than 200 upvotes. Your next target is to reach 300 upvotes.

You can view your badges on your board And compare to others on the Ranking
If you no longer want to receive notifications, reply to this comment with the word STOP

To support your work, I also upvoted your post!

Do not miss the last post from @hivebuzz:

Feedback from the last Hive Power Up Day