Short-Term Meaning Shift: A Distributional Exploration

TL;DR

This paper investigates short-term meaning shifts in online communities using distributional models, identifies challenges in distinguishing shifts from referential changes, and proposes a new measure to improve detection accuracy.

Contribution

It introduces the first dataset and analysis of short-term meaning shift, highlighting model limitations and proposing a novel measure of contextual variability.

Findings

Model struggles to differentiate meaning shift from referential phenomena.

Proposed measure of contextual variability improves detection.

First annotated dataset for short-term meaning shift in online communities.

Abstract

We present the first exploration of meaning shift over short periods of time in online communities using distributional representations. We create a small annotated dataset and use it to assess the performance of a standard model for meaning shift detection on short-term meaning shift. We find that the model has problems distinguishing meaning shift from referential phenomena, and propose a measure of contextual variability to remedy this.