Yang Liu

Yang Liu is a researcher at the Institute of Computing Technology (ICT), Tsinghua University, and affiliated with the Beijing Academy of Artificial Intelligence. His work spans computational linguistics, machine translation, and dialogue systems, with additional collaborations in medical NLP and cross-lingual knowledge representation.

• Doctorate from Purdue University
• Collaborations with Microsoft, Amazon, and Facebook researchers

Research focuses on Neural Machine Translation (NMT) advancements through:

• Self-supervised alignment and regularization
• Document-level context integration
• Robust translation frameworks for low-resource scenarios
• Adapter-based pluggable models

Recent works extend to conversational AI, including:

• Subjective knowledge integration in task-oriented dialogue
• Context representation analysis for open-domain dialogue
• Synthetic conversation dataset generation

Key contributions include:

• THUMT: Open-source NMT toolkit
• Earth Mover's Distance for lexicon induction
• Context Gates for translation fluency

Publications demonstrate continuous engagement with top-tier conferences including ACL, EMNLP, and COLING, collaborating with prominent figures like Maosong Sun and Qun Liu.