Just-DREAM-about-it: Figurative Language Understanding with DREAM-FLUTE

TL;DR

This paper introduces DREAM-FLUTE, a system that enhances figurative language understanding by constructing mental models of scenes, leading to improved performance and interpretability in language tasks.

Contribution

It presents a novel approach combining scene elaboration with language understanding, demonstrating significant performance gains in figurative language tasks.

Findings

Achieved first place in FigLang2022 Shared Task with 63.3% accuracy.

Adding a mental model improves accuracy by 3.3%.

Ensemble techniques further enhance performance.

Abstract

Figurative language (e.g., "he flew like the wind") is challenging to understand, as it is hard to tell what implicit information is being conveyed from the surface form alone. We hypothesize that to perform this task well, the reader needs to mentally elaborate the scene being described to identify a sensible meaning of the language. We present DREAM-FLUTE, a figurative language understanding system that does this, first forming a "mental model" of situations described in a premise and hypothesis before making an entailment/contradiction decision and generating an explanation. DREAM-FLUTE uses an existing scene elaboration model, DREAM, for constructing its "mental model." In the FigLang2022 Shared Task evaluation, DREAM-FLUTE achieved (joint) first place (Acc@60=63.3%), and can perform even better with ensemble techniques, demonstrating the effectiveness of this approach. More generally, this work suggests that adding a reflective component to pretrained language models can improve their performance beyond standard fine-tuning (3.3% improvement in Acc@60).