Kewei Tu
Scholar

Kewei Tu

Google Scholar ID: 5gi3Pm0AAAAJ
School of Information Science and Technology, ShanghaiTech University, China
Natural Language ProcessingMachine Learning
Citations & Impact
All-time
Citations
2,236
 
H-index
25
 
i10-index
54
 
Publications
20
 
Co-authors
14
list available
Resume (English only)
Academic Achievements
  • Published 'Parallel Continuous Chain-of-Thought with Jacobi Iteration' at EMNLP 2025.
  • Published 'Tree-Structured Non-Autoregressive Decoding for Sequence-to-Sequence Text Generation' in Findings of EMNLP 2025.
  • Published two papers at ACL 2025: 'A Systematic Study of Compositional Syntactic Transformer Language Models' and 'Look Both Ways and No Sink: Converting LLMs into Text Encoders without Training'.
  • Published 'A Systematic Study of Cross-Layer KV Sharing for Efficient LLM Inference' at NAACL 2025.
  • Published three papers at ACL 2024, including 'Layer-Condensed KV Cache for Efficient Inference of Large Language Models', 'Dependency Transformer Grammars', and 'Generative Pretrained Structured Transformers'.
  • Published 'Augmenting transformers with recursively composed multi-grained representations' at ICLR 2024.
  • Published 'Frame Semantic Role Labeling Using Arbitrary-Order Conditional Random Fields' at AAAI 2024.