Retrieved evidence supports the claim that round-trip translation serves as an effective technique for machine translation quality evaluation, provided methodological artifacts like copying mechanisms are controlled for.
Papers [4] and [9] explicitly utilize and validate round-trip translation as an evaluation technique for machine translation performance and quality analysis. Other papers discuss round-trip translation for data augmentation or training (e.g., [0], [1], [2], [6], [10]), but papers [4] and [9] directly address its efficacy as an evaluation technique, supporting the claim.