Yesterday I was looking at Cindicator, which combines the intelligence of millions of financial analysts with the power of AI. Today it’s Numerai, which is doing something similar.
Numerai is incentivizing the construction of an artificial intelligence hedge fund with Numeraire, a cryptographic token on the Ethereum blockchain. It’s an AI-run, crowd-sourced hedge fund.
By doing this Numerai is bringing network effects and collaboration to finance.
Numerai believes that the stock market is inefficient with respect to new developments in machine learning. Only a fraction of the world’s data scientist has access to its data. Numerai wants to change that.
It abstracts its financial data so that data scientists do not know what the data represents. Doing this removes human biases and overfitting.
Overfitting is a modeling error that occurs when a function is too closely fit to a limited set of data points. The result is an overly complex model that’s unlikely to work well with new data.
Here’s how it works. Data scientists become part of the ecosystem by competing in tournaments.
Analysts get encrypted and contextless data to avoid human biases. The analysts then process the data and build their own models. And they bet NMR tokens to support their predictions. The more they bet, the higher their confidence.
They get rewarded in dollars from the weekly payout pool in proportion to their stake. The staked NMR gets returned. When predictions turn out wrong, their NMR stake is destroyed.
The models submitted by data scientists are synthesized by Numerai into a meta model. So Numerai gets exposure to every participating analyst’s model.
Numerai then uses the metal model and AI to run its hedge fund.
According to CEO Richard Craib it already has 30,000 data scientists who have uploaded over 1 million prediction sets. Read about his master plan here.
And check out the whitepaper here.
Looks pretty cool. Some big name backers. Low market cap. What do you think of Numeraire?