Token Portfolio Optimizing Algorithm in Action

Welcome back!

In this video/post, we’ll talk a bit more about what TokenAI is doing in terms of AI.

In the video above, you can see a working example of our portfolio optimizing algorithm which is part of our AI Reinforcement Learning Module.

Disclaimer: this algorithm is still in beta and that this shouldn’t be considered as investment advice in any kind.

What you are looking at in the video is the TokenAI algorithm running through actual currency market data that spans over 65 weeks or so. The data includes token performance, market cap and other attributes that allow the algorithm to look for trends and identify breakouts in currency performance.

As we show in the video, the algorithm not only can deliver incredible performance, it also beats a market portfolio that doesn’t use AI to rebalance it.

I also want to share some of what we are working on in terms of AI (this is an excerpt from our white paper that can be found at www.tokenai.io)

The TokenAI platform is a set of AI modules that can be selected and combined to produce a cryptocurrency portfolio or to analyze and optimize an existing portfolio.

We think of the modules as individual AIs — each continuously learning from different sources of data — and all at the user’s disposal.

The AI Reinforcement Learning Module is the AI algorithm that manages the TokenAI basket index. It uses its own change in valuation as a reward and punishment signal to adjust the basket index cryptocurrency portfolio. Over time this module will learn through monitoring what the aggregate market thinks of its performance how to perform better.

The AI Anomalous Movement Module quantifies typical movement across all of the cryptocurrencies. This module has the ability to both alert the user of the movement and autonomously construct trades based on the detected anomalies.

The AI Mimicking Module allows users to train the AI to mimic and adapt to their own trading preferences allowing them to automate and improve trading. This module watches trades the user makes and as it observes trading habits it will begin suggesting similar trades. Once a sufficient level of trust is developed, Users can then allow it to place the trades for them.

The AI Social Module uses sentiment analysis from news and discussion on major news sites, blogs, and social media. This module will give the user deep analysis of the global sentiment (even from sites in many languages that the users can’t even read!) and provide the ability to construct trades based on trends and influencers invisible to most users.

The AI Risk Optimization Module uses the historical data from the cryptocurrencies to quantify and optimize the amount of risk the user is willing to accept. With the immaturity and lack of sophistication of many traders in these markets, having well quantified risk profile is crucial to long-term success.

The AI Market Manipulation Avoidance Module assumes that there will continue to be malicious actors who attempt to both manipulate the currencies’ prices and the exchanges’ ability to operate. By monitoring activity across all of the exchanges, this module provides users and exchanges with analysis of how well it believes true price discovery is working and prevents them from being cheated.

I hope you enjoyed this post. We will be posting more cool videos of what we’re doing and how our AI algorithms work on this blog.

Also, please check out www.tokenai.io for more information.