Comparable Corpora: Opportunities for New Research Directions
Kenneth Church
TL;DR
This paper surveys Comparable Corpora (CC) as a rich resource beyond traditional parallel corpora, arguing for broader research directions. It traces CC's historical lineage from parallel resources to CC, discusses word-sense disambiguation, lexical semantics, and BLI, and highlights translation artifacts that limit current benchmarks. Key proposals include avoiding pivoting via English, leveraging transfer learning to growth languages, embracing multimodal CC, and addressing filter-bubble biases in bots and news. The work calls for new benchmarks, richer cross-linguistic lexical representations, and interdisciplinary methods to exploit CC for deeper cross-lingual understanding and inclusive AI systems.
Abstract
Most conference papers present new results, but this paper will focus more on opportunities for the audience to make their own contributions. This paper is intended to challenge the community to think more broadly about what we can do with comparable corpora. We will start with a review of the history, and then suggest new directions for future research. This was a keynote at BUCC-2025, a workshop associated with Coling-2025.