text-segmentation-three-distinct-tasks-taxonomy

Summary

In standard NLP, "text segmentation" is an umbrella label that covers three genuinely different jobs: finding where words start and end, finding where one sentence ends and the next begins, and grouping words into larger phrases or clauses. This matters because a system that assumes these are interchangeable will misapply methods or evaluation metrics — the three tasks operate at different scales of text and require different techniques, so conflating them under one label can lead to sloppy task specification.