/CEM

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.

Primary LanguagePythonMIT LicenseMIT

Watchers