Search papers, labs, and topics across Lattice.
This study critically evaluates the validity of knowledge bases (KBs) as gold standards in entity-level machine translation, particularly focusing on the translation of person names in low-resource historical texts. By comparing expert annotations from the National Institute of Korean History against KB-injected translations, the authors reveal a significant self-referential bias in evaluation metrics, where 97.8% agreement is observed in the overlapping segment, yet the independent segment shows no improvement. The findings indicate that the perceived performance gains from KB injection are misleading, as they are confined to segments benefiting from the same resource, undermining the reliability of KBs in assessing translation quality.
The reliance on knowledge bases as gold standards in machine translation may inflate performance metrics, masking the true quality of translations in low-resource settings.
The Seungjeongwon Ilgi, a UNESCO Memory of the World record, is only 37.4% translated, and the most conspicuous failure mode in automatic translation is the person name -- a misread name corrupts the historical fact rather than merely the surface. Low-resource historical domains have no expert gold standard for entity translation, so practitioners substitute a knowledge base (KB) for the gold. That KB is the same resource injected into the system: scoring becomes self-referential and the metric measures instruction compliance rather than translation quality. We measure this loop. Using expert person-name annotations from the National Institute of Korean History as a gold independent of the injection pipeline, we hold the entity set fixed and vary only the provenance of the correct reading. Of 527 expert-annotated mentions, only 31.1% lie outside the injection pipeline, and the residual loop is not uniform -- in the overlapping segment the injected reading agrees with the human translation 97.8% of the time against 70.1% in the independent one, so the segment that looks healthiest is the one the loop is holding up. Across four models, a difference-in-differences analysis shows the gain from KB injection is confined to the segment whose gold shares the injected resource; in the independent segment it is at or below zero. Post-injection preservation clusters in a narrow 0.910-0.996 band even though baseline capability differs fivefold, so the reported gain is the complement of prior performance and weaker models appear to improve more dramatically. On an independent sample built by removing the construction filter, the measure replicates within model (overlapping intervals) while discriminating between models (non-overlapping intervals) -- it reflects a property of the model, not of the sample.
