CharManteau: Character Embedding Models For Portmanteau Creation

TL;DR

This paper introduces character-level neural sequence-to-sequence models for portmanteau creation, demonstrating improved accuracy and human evaluation results over traditional methods, without relying on phonetic information.

Contribution

It presents a novel noisy-channel-style neural model for portmanteau generation that leverages unsupervised word lists and exhaustive candidate generation.

Findings

Outperforms FST-based baseline in accuracy

Achieves higher human evaluation scores

Effective without phonetic information

Abstract

Portmanteaus are a word formation phenomenon where two words are combined to form a new word. We propose character-level neural sequence-to-sequence (S2S) methods for the task of portmanteau generation that are end-to-end-trainable, language independent, and do not explicitly use additional phonetic information. We propose a noisy-channel-style model, which allows for the incorporation of unsupervised word lists, improving performance over a standard source-to-target model. This model is made possible by an exhaustive candidate generation strategy specifically enabled by the features of the portmanteau task. Experiments find our approach superior to a state-of-the-art FST-based baseline with respect to ground truth accuracy and human evaluation.