ReCo: Reliable Causal Chain Reasoning via Structural Causal Recurrent Neural Networks

TL;DR

ReCo introduces a structural causal recurrent neural network framework that enhances causal chain reasoning by modeling threshold and scene contradictions, improving reliability and downstream task performance.

Contribution

The paper proposes ReCo, a novel framework that incorporates exogenous variables and SRNN to address transitive issues in causal chain reasoning, advancing the reliability of AI decision-making.

Findings

ReCo outperforms strong baselines on CCR datasets in Chinese and English.

Injecting ReCo's causal chain knowledge improves BERT's performance on causal tasks.

ReCo effectively models threshold and scene contradictions in causal reasoning.

Abstract

Causal chain reasoning (CCR) is an essential ability for many decision-making AI systems, which requires the model to build reliable causal chains by connecting causal pairs. However, CCR suffers from two main transitive problems: threshold effect and scene drift. In other words, the causal pairs to be spliced may have a conflicting threshold boundary or scenario. To address these issues, we propose a novel Reliable Causal chain reasoning framework~(ReCo), which introduces exogenous variables to represent the threshold and scene factors of each causal pair within the causal chain, and estimates the threshold and scene contradictions across exogenous variables via structural causal recurrent neural networks~(SRNN). Experiments show that ReCo outperforms a series of strong baselines on both Chinese and English CCR datasets. Moreover, by injecting reliable causal chain knowledge distilled by ReCo, BERT can achieve better performances on four downstream causal-related tasks than BERT models enhanced by other kinds of knowledge.