The softmax function appears in many machine learning algorithms, in particular neural networks and prediction markets. The idea is, if you have a set of values, to scale them so they sum to 1.0 and therefore can be interpreted as … Continue reading The Max Trick when Computing Softmax