4 ms·
I wonder what classifier was used (assume neural network-based, given the figures), and how that compares to a simple baseline that uses bag-of-words, such as a
by blackbear_ 6y ago
I wonder what classifier was used (assume neural network-based, given the figures), and how that compares to a simple baseline that uses bag-of-words, such as a linear model or naive Bayes. The examples look easy enough to be classified by matching keywords.
- rezamoaiandin 6y agoWe used a shallow neural network. The main challenge at Sculpt wasn't the modelling part, but rather the whole UX (active learning to speed up the training process, explainability, easily get predictions, etc). So it's true that a relatively simple model performs well, and actually having an efficient / shallow network also helps make the active learning pipeline fast for the user.
- blackbear_ 6y agoThanks! The dashboard looks like a nice piece of work indeed.