Joint intent detection and slot filling using weighted finite state

Task Conditioned BERT for Joint Intent Detection and Slot-Filling bert slot Experimental results demonstrate that our proposed model achieves significant improvement on intent classification accuracy, slot filling F1, bert slot Predict intent and slot at the same time from one BERT model ; total_loss = intent_loss + coef * slot_loss (Change coef with --slot_loss_coef

osrs ranged head slot Experimental results demonstrate that our proposed model achieves significant improvement on intent classification accuracy, slot filling F1,