Guided Generation of Cause and Effect

TL;DR

This paper introduces a new framework for generating cause-and-effect sentences using large causal resources and improved decoding, achieving better causal reasoning performance.

Contribution

It develops large-scale causal sentence collections and enhances lexically-constrained decoding for high-quality causal text generation.

Findings

High-quality, diverse causal sentence generation confirmed by human assessment.

3-point improvement on COPA causal reasoning benchmark without changing model architecture.

Introduction of CausalBank and refined Cause Effect Graph resources.

Abstract

We present a conditional text generation framework that posits sentential expressions of possible causes and effects. This framework depends on two novel resources we develop in the course of this work: a very large-scale collection of English sentences expressing causal patterns CausalBank; and a refinement over previous work on constructing large lexical causal knowledge graphs Cause Effect Graph. Further, we extend prior work in lexically-constrained decoding to support disjunctive positive constraints. Human assessment confirms that our approach gives high-quality and diverse outputs. Finally, we use CausalBank to perform continued training of an encoder supporting a recent state-of-the-art model for causal reasoning, leading to a 3-point improvement on the COPA challenge set, with no change in model architecture.