Reproduce “HERIT: Enabling Global Access to Hanja-to-English Historical Translation by Mitigating Temporal Bias” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Dayeon Jeong, Kyeongpil Kang
HERIT translates historical Hanja records into English using existing Korean translations and expert examples to create pseudo-labeled training data, then fine-tunes on human-translated material. In a 200-passage held-out evaluation, experts rated it 3.93 out of 5, versus 3.22 for Qwen3-32B trained only on human data. This supports gains on tested AJD periods, not verified quality across AJD and JRS outputs.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.