Interesting, was not familiar with this concept — and I indeed see the similarity here, the model learns to stray from risk. Nice.Will be interesting to train a risk-loving model (minimize entropy) and see how it performs on different challenges..