Scholar
Zheng Wei Lim
Google Scholar ID: VRBbHDcAAAAJ
The University of Melbourne
Citations & Impact
Citations
78
H-index
3
i10-index
2
Publications
8
Co-authors
3
list available
Publications
Resume (English only)
Academic Achievements
- Selected Publications:
- [1] Language-Specific Latent Process Hinders Cross-Lingual Performance
- [2] Mufu: Multilingual Fused Learning for Low-Resource Translation with LLM
- [3] Simpson’s Paradox and the Accuracy-Fluency Tradeoff in Translation
- [4] A Computational Approach to Identifying Cultural Keywords Across Languages
- [5] Predicting Human Translation Difficulty with Neural Machine Translation
- [6] Predicting Human Translation Difficulty Using Automatic Word Alignment
Research Experience
- In 2024, a student researcher at Google Research Australia. Mainly working on computational methods for difficult (i.e., cross-cultural or low-resource) translation, multilinguality, psycholinguistics, and interpretability.
Education
- PhD student in NLP at the University of Melbourne, advised by Ekaterina Vylomova, Charles Kemp, and Trevor Cohn
Background
- Research interests include understanding how model architecture, training data, and algorithms impose learning biases on language models, and their limitations in representing cognitively driven language phenomena; studying how large-scale models navigate varying (sometimes conflicting) goals, and developing controllable mechanisms that drive convergent or adaptable behavior across time, domains, languages, and modalities.
