Scholar
Zhengyang Tang
Google Scholar ID: 2RRV0PQAAAAJ
CUHKSZ
Citations & Impact
Citations
3,963
H-index
8
i10-index
8
Publications
13
Co-authors
9
list available
Publications
Resume (English only)
Academic Achievements
- NeurIPS 2025: CoRT (Code-integrated Reasoning within Thinking)
- COLM 2025: SCRIT (Self-Evolving Critique Abilities in LLMs)
- ICML 2024: MathScale (Scaling Instruction Tuning for Mathematical Reasoning)
- TMLR 2025: GLAN (Generalized Instruction Tuning for Language Models)
- ACL 2025: Second Language (Arabic) Acquisition of LLMs via Progressive Vocabulary Expansion (Oral & Panel)
- COLING 2022: DPTDR (Deep Prompt Tuning for Dense Passage Retrieval)
- Operations Research 2025: ORLM (Customizable Framework for Automated Optimization Modeling)
- Contributed to Qwen3 Technical Report (Tool-integrated Reasoning)
Background
- Ph.D. candidate at The Chinese University of Hong Kong, Shenzhen, advised by Prof. Benyou Wang.
- Research focuses on developing intelligent agents capable of complex reasoning and self-improvement.
- Pioneers agentic frameworks leveraging reinforcement learning (RL) for tool-integrated tasks.
- Proposed SCRIT, a self-evolving critique model serving as a generative reward model for scalable, supervision-free oversight.
- Designed novel instruction tuning frameworks—MathScale, GLAN, and ALAN—for scalable high-quality training data generation.
- Also works on efficient information access methods (e.g., DPTDR), achieving top performance on competitive benchmarks.
