BottleHumor: Self-Informed Humor Explanation using the Information Bottleneck Principle

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Hwang, EunJeong, West, Peter, Shwartz, Vered
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916986144948224
author Hwang, EunJeong
West, Peter
Shwartz, Vered
author_facet Hwang, EunJeong
West, Peter
Shwartz, Vered
contents Humor is prevalent in online communications and it often relies on more than one modality (e.g., cartoons and memes). Interpreting humor in multimodal settings requires drawing on diverse types of knowledge, including metaphorical, sociocultural, and commonsense knowledge. However, identifying the most useful knowledge remains an open question. We introduce \method{}, a method inspired by the information bottleneck principle that elicits relevant world knowledge from vision and language models which is iteratively refined for generating an explanation of the humor in an unsupervised manner. Our experiments on three datasets confirm the advantage of our method over a range of baselines. Our method can further be adapted in the future for additional tasks that can benefit from eliciting and conditioning on relevant world knowledge and open new research avenues in this direction.
format Preprint
id arxiv_https___arxiv_org_abs_2502_18331
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle BottleHumor: Self-Informed Humor Explanation using the Information Bottleneck Principle
Hwang, EunJeong
West, Peter
Shwartz, Vered
Computation and Language
Humor is prevalent in online communications and it often relies on more than one modality (e.g., cartoons and memes). Interpreting humor in multimodal settings requires drawing on diverse types of knowledge, including metaphorical, sociocultural, and commonsense knowledge. However, identifying the most useful knowledge remains an open question. We introduce \method{}, a method inspired by the information bottleneck principle that elicits relevant world knowledge from vision and language models which is iteratively refined for generating an explanation of the humor in an unsupervised manner. Our experiments on three datasets confirm the advantage of our method over a range of baselines. Our method can further be adapted in the future for additional tasks that can benefit from eliciting and conditioning on relevant world knowledge and open new research avenues in this direction.
title BottleHumor: Self-Informed Humor Explanation using the Information Bottleneck Principle
topic Computation and Language
url https://arxiv.org/abs/2502.18331