dettnet

dettnet, using highly compact ai model and the benefits of fast processing blockchain networks, we can deploying ai model on chain. we are testing it with decision tree model, because the architecture of our ttnet allows us to turn any model into rule based model in the end of the training.

we have our source code for the model pipeline in the published paper, and my job in this hackathon is to test the concept of deploying decision tree model on chain, and test it is fast enough for processing some ai inference. in the end, we compare the inference cost in terms of money and time roughly in different deployment environment.