Cross-Register Projection for Headline Part of Speech Tagging
This work addresses the specific issue of improving NLP models for news headlines, which is incremental as it adapts existing methods to a domain-specific register.
The paper tackled the problem of part-of-speech tagging for English news headlines, which underperform due to register differences from long-form text, by projecting tags from news bodies and training a multi-domain tagger, achieving a 23% relative error reduction per token and 19% per headline on a new corpus of 5,248 headlines.
Part of speech (POS) tagging is a familiar NLP task. State of the art taggers routinely achieve token-level accuracies of over 97% on news body text, evidence that the problem is well understood. However, the register of English news headlines, "headlinese", is very different from the register of long-form text, causing POS tagging models to underperform on headlines. In this work, we automatically annotate news headlines with POS tags by projecting predicted tags from corresponding sentences in news bodies. We train a multi-domain POS tagger on both long-form and headline text and show that joint training on both registers improves over training on just one or naively concatenating training sets. We evaluate on a newly-annotated corpus of over 5,248 English news headlines from the Google sentence compression corpus, and show that our model yields a 23% relative error reduction per token and 19% per headline. In addition, we demonstrate that better headline POS tags can improve the performance of a syntax-based open information extraction system. We make POSH, the POS-tagged Headline corpus, available to encourage research in improved NLP models for news headlines.