ATD-Trans: A Geographically Grounded Japanese-English Travelogue Translation Dataset

2026-05-12

AI Summary

This study addresses the lack of fine-grained, geographically grounded evaluation benchmarks for machine translation of multilingual geographic texts, such as travelogues. Focusing on the Japanese–English language pair, we introduce ATD-Trans, the first parallel travelogue corpus annotated with geographic entities. This dataset enables translation quality assessment at both the document and geographic entity levels, while distinguishing performance on domestic (Japan-based) versus international locations. Building upon this resource, we propose an evaluation framework that integrates geographic entity recognition and cross-lingual alignment. Comparative experiments reveal that models optimized for Japanese achieve superior overall performance, yet struggle more with accurately translating domestic geographic entities. Our work establishes a foundational benchmark for multilingual geographic text processing and offers a novel paradigm for fine-grained cross-lingual geographic understanding.

Abstract

Geographic text, or textual data rich in geographic (geo-) information is a valuable source for various geographic applications, e.g., tourism management. Making such information accessible to speakers of other languages further enhances its utility; thus, accurate machine translation (MT) is essential for equity in multilingual geo-information access. To facilitate in-depth analysis for geographic text, we introduce ATD-Trans, a geographically grounded Japanese--English travelogue translation dataset, which enables evaluation of MT quality at both the overall and geo-entity levels across domestic (within Japan) and overseas regions. Our experiments on existing language models examine two factors: model language focus and geographic regions. The results highlight advantages of Japanese-enhanced models and greater difficulty in translating domestic-region geo-entities mentioned in travel blogs.

Problem

• geographic text
• machine translation
• Japanese-English translation
• geo-entity
• travelogue

Innovation

• geographically grounded translation
• Japanese-English travelogue dataset
• geo-entity translation
• machine translation evaluation
• multilingual geo-information