Working Across Languages in NLP

Rachel Wagner-Kaiser, Tim Cerino · 2025

This chapter covers the challenges in working across languages, starting with the unique challenges of optical character recognition (OCR) and speech to text (STT) for non-English languages. Techniques and options for improving extraction of text from images and audio is covered, with a focus on the challenges with non-English languages. The chapter also reviews options and approaches for working across multiple languages, additional processing techniques and considerations, the role of LLMs, and the factors to be considered in the decision of whether to build a solution for each language or a cross-language solution.

Read the paper · More papers on PaperTik