In 2020, Permutable solved a one-shot learning classification problem of recognising pattern formations.  Even to this day, modern charting services do not have such capability.  The below is a result of deploying real-time and light-weight machine learning models across a web server cloud matrix.

R2 uses state-of-the-art ‘one-shot’ and ‘small sample’ machine learning techniques to provide trading insights and entry points across wide array of assets. Opportunities are provided to the user via a interactive heat map which be configured to the user’s demands.

Model training can be undertaken via two methods. The AI can be trained by selecting patterns on historical chart data or by creating a nodal representation of the pattern.