EMNLP'22, CEM improves MHCH performance by correcting prediction bias and training an auxiliary cost simulator based on user state and labor cost causal graph, without requiring complex model crafting.
python
natural-language-processing
tensorflow
causal-inference
dialogue-systems
counterfactual
causal-graph
-
Updated
Oct 9, 2022 - Python