Just-DREAM-about-it: Figurative Language Understanding with DREAM-FLUTE
Gu, Yuling, Fu, Yao, Pyatkin, Valentina, Magnusson, Ian, Mishra, Bhavana Dalvi, Clark, Peter
–arXiv.org Artificial Intelligence
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." Figure 1: Overview of DREAM-FLUTE: It first uses In the FigLang2022 Shared Task evaluation, DREAM (Gu et al., 2022) to generate an elaboration of DREAM-FLUTE achieved (joint) first place the situation in the premise and hypothesis (separately), (Acc@60=63.3%), and can perform even better then uses this additional context for entailment classification with ensemble techniques, demonstrating and explanation generation.
arXiv.org Artificial Intelligence
Oct-28-2022
- Country:
- North America > United States
- Washington > King County
- Seattle (0.04)
- Minnesota > Hennepin County
- Minneapolis (0.04)
- Washington > King County
- Europe > Ireland
- Leinster > County Dublin > Dublin (0.04)
- Asia > China
- Hong Kong (0.04)
- Africa > Ethiopia
- Addis Ababa > Addis Ababa (0.04)
- North America > United States
- Genre:
- Research Report (0.40)
- Technology: