π¦Institutions / Developer
HiveAI (using Swarm Intelligence for Market Predictions), API Predictions and Data Access, and the 3rd Bot Market Place for Trading Strategies.
Crypticorn provides special access and capabilities for developers, trading algorithm creators, and institutions.
Build AI models and earn $AIC
Get API access to live data for your AI model
Get API access to predictions data to build your trading algorithm
Create and earn for your trading bots through the 3rd Bot Marketplace
API Access: Data and Predictions
Get access to Crypticorn's data for developing your own models and prediction data to create your own trading bots.
Data Access:
We mine a vast amount of data, specifically preparing it for AI tasks. Everyone can access this data and train their own models for HiveAI or personal use through our API. We specifically design and prepare this data for AI tasks, and you can easily use that.
Predictions Access:
On the other side, the predictions made by Crypticornβs AI and HiveAI can be accessed by trading bots for personal usage, institutions for their trading, or developers of the third bot marketplace.
HiveAI (Decentralised AI)
HiveAI leverages the swarm intelligence of thousands of people to create the best predictions for cryptocurrencies.
For Developers:
π°Getting started as Developer$AIC Usage:
To get more rewards, stake AIC for successful models.
Pay in AIC for live data so that your model can use it as you wish (API access).
Problem:
Before you can even begin AI selection, training, and other processes, data mining and data preparation are the biggest and often most time-consuming steps. Getting all the historical data, treating missing values, and bringing them into shape is a very hard and annoying task.
Especially when you want non-trivial data like Google Trends, you need proxy rotation to gather this special type of data, which requires weeks to implement properly before you can even test if this data is worth the effort. Itβs even worse for live predictions.
Furthermore, the process of preparing data, creating features, setting targets, and setting lags is also the most prone to errors, often leading to look-ahead bias or data leakage.
After mining and preparing the data and targets, we have the next problem: the vast amount of potential AI models (tree-based, ARIMA-family, LSTMs, GARCH, ensembles, stacking, bagging, and more machine and deep learning frameworks) to predict the many different targets (15 minutes, next day close, next day volatility, next day market regimes, next day up or down, etc.).
Therefore, we need to adjust the model's feature importance, cross-validation, and structure based on the specific targets or underlying asset. A small team could never test all those possibilities in a short amount of time to find the most promising solutions and adapt it to the different targets and assets.
Solution/Idea:
Crypticorn has already mined a lot of data and will continue to gather and collect more, especially special data that is not easily available and requires a lot of effort to obtain
Furthermore, we use paid API sources that a single person or a small group might not want to pay for. Therefore, we could make this data available to everyone, including those who are familiar with AI models for time series prediction and classification tasks.
We would obfuscate the feature names and target names to prevent look-ahead bias, bias in general, an understanding of the underlying asset, and more (see example below).
Crypticorn would then standardize this dataset, split it into a test and training dataset, and only distribute the X_train, Y_train, and X_test. We would not give out the y_test data set.
Anyone who desires to participate in this challenge can attempt to train AI models using this data, generate predictions for the X_test, and then forward these predictions to our API for data evaluation.
If the creator's models continue to perform exceptionally well under real-world conditions, they can submit them to Crypticorn for access, profit-sharing, AIC rewards, and other rewards based on those evaluation results.
Those predictions wonβt be used directly. Crypticorn will combine all these models to form an ensemble that is suitable for various targets and assets. It then generates the final forecasts for the prediction dashboard or AI trading bots.
Advantages:
Using the swarm intelligence of thousands of people, we can test thousands of different approaches.
These developers donβt need to understand cryptocurrencies, trading, etc.
The participants will/could train the models locally. No additional resources at Crypticorn needed.
Events to expand Crypticorn's reach (Kaggle events, tournaments and more).
Stay ahead of everyone else, as we will automatically get new models, new approaches, and more.
For the final predictions, we only need a small team inside Crypticorn to maintain the meta-model AI.
Faster AI development
It doesnβt conflict with Crypticornβs current actions.
We set the targets in such a way that they are the most useful for our trading strategies.
3rd Bot Marketplace
The 3rd Bot Marketplace is the integration of other developers, quants, and institutionsβ trading bots into the Crypticorn ecosystem. This trading bot utilizes the Crypticorn data pipeline, predictions, and trade execution systems. This makes it easy for the developers to create strategies and then allow them to trade.
Crypticorn will test and evaluate these trading strategies before token holders can select them and allocate funds to make them. This gives all Crypticorn token holders access to a wider range of strategies.
The performance fee model also accounts for these bots, and the developer decides on the height of the performance fees from the user. Depending on the trading strategy's complexity and resource usage, the developer earns B% of the performance fees, while the remaining A% of the performance fees go to Crypticorn for the use of our technology and trade execution.
Last updated