Logical Inference for Numerical Comparatives between Semi-structured Tables and Texts

Abstract

The Natural Language Inference (NLI) task has recently been studied for inferences between semi-structured tables and texts. Although neural approaches have achieved high performance in various types of NLI, including NLI between semi-structured tables and texts, they still have difficulty in performing a numerical type of inference, such as counting. We propose a logical inference system to handle a numerical type of inference between semi-structured tables and texts. We use logical representations as meaning representations for tables and texts and use model checking. To evaluate the extent to which our system can perform inference with numerical comparatives, we provide a test set and an evaluation protocol for a numerical type of inference between semi-structured tables and texts in English. We show that our system can more robustly perform inference between tables and texts that requires numerical understanding compared with current neural approaches.