Decomposing Natural Logic Inferences in Neural NLI

Anonymous

Abstract: In the interest of interpreting neural NLI models and their reasoning strategies, we carry out a systematic probing study which investigates whether these modelscapture the crucial semantic features central to natural logic: \emph{monotonicity} and \emph{concept inclusion}.Correctly identifying valid inferences in \emph{downward-monotone contexts} is a known stumbling block for NLI performance,subsuming linguistic phenomena such as negation scope and generalized quantifiers.To understand this difficulty, we emphasize monotonicity as a property of a \emph{context} and examine the extent to which models capture monotonicity information in the contextual embeddings which are intermediate to their decision making process.Drawing on the recent advancement of the probing paradigm,we compare the presence of monotonicity features across various models.We find that monotonicity information is notably weak in the representations of popularNLI models which achieve high scores on benchmarks, and observe that previous improvements to these models based on fine-tuning strategies have introduced stronger monotonicity features together with their improved performance on challenge sets.