Sun Xiaoxi 孙小翕

Education:
B.A., Beijing Foreign Studies University
MTI, Renmin University of China

Research Interest:
Translation and Interpreting Assessment; Cognitive‑process Studies; Translation Quality Metrics; Human Evaluation of Machine Translation.

Self-Introduction:
My work centers on human judgement in translation evaluation. I am interested in how evaluators form quality judgements for machine‑generated translations, and what factors lead to inconsistent ratings among different raters.

Thesis Advisor: A/P Han Chao

Thesis Title:  Modeling Rater Cognition in MQM-Style Assessment of Machine Translation (tentative)

Thesis Research Area:
Human evaluation of machine translation provides a vital basis for ensuring translation quality and developing automatic evaluation metrics. Even with fine‑grained quality measurement tools such as Multidimensional Quality Metrics (MQM), low inter‑annotator agreement persists within human annotation practice. Existing studies predominantly focus on evaluation outputs, with limited attention devoted to raters’ underlying cognitive processing. Relying on empirical experiments and self‑built domain‑specific corpora, this study investigates how multiple conditions jointly influence raters’ annotation behaviors and explores the cognitive traces of rating divergences. It aims to provide references for rater training and the optimization of translation‑evaluation systems.