There is no fixed schedule for VAI’s model updating. It depends on the arrival of new datasets as well as market activity.
We apply a well known technique in machine learning to monitor model performance called model drift. This measure triggers new model training, depending on how much the model has drifted away from the market since the last update.