🔮Fama's AI

We use a special type of AI called Reinforcement Learning (RL) to make predictions on the future price of crypto assets. RL has historically been used in video games to train computer controlled players and rewards the AI for interacting positively with its environment. In Fama, we take the same concept and use the crypto market as the environment in order to train the model! The RL model is given years of data about the assets its being designed to trade. During this training period, the RL model is trying build up a reference of how the data throughout the years relates to eachother. The nature of this type of technology means we don't truly know what kind of connections and patterns the model is forming and is likely beyond our comprehension or would take us a long time to figure out. Once complete, the RL is fed fresh information and asked to generate predictions about the outcome. We interpret those predictions and convert them into "buy", "sell" or "hold". The Fama Whitepaper (Section 5) has a great technical overview of the process, data and technology that we are exploring in order to create our AI and signals.