Zecheng Tang (汤泽成)

---

---

---

Brief Bio

I am currently a researcher at Moonshot AI, and concurrently a final-year PhD candidate at Soochow University (exp. grad. Jul 2027), supervised by Prof. Juntao Li and Prof. Min Zhang. My research focuses on machine learning and long-sequence modeling (with Algorithm-Architecture co-design). My contributions include but not limited to: (1) Foundation Model: Kimi-K3, Qwen-Image-series: ( Qwen-Image / Qwen-Image-2.0 / Qwen-Image-2.0-RL ) and OpenBA-series: ( OpenBA-V1 / OpenBA-V2 ); (2) AIGC: Visual ChatGPT, NUWA-series: ( LayoutNUWA / StrokeNUWA ).

Before that, I obtained my Bachelor's degree in Software Engineering from Soochow University (ranked 1st / 380). My formal name is Zecheng Tang (汤泽成), and you may call me Zecheng (/zə'tʃɛŋ/), or simply Tang.

---

Work Experience

---

Selected Conference & Journal

* indicates equal contribution; Full list: Google Scholar; Rank system: CCF (China Computer Federation)

[Click to hide / view selected paper list]

Long Sequence Modeling (Optimization)

• Elastic Attention: Test-time Adaptive Sparsity Ratios for Efficient Transformers
  Zecheng Tang, Quantong Qiu, Yi Yang, Zhiyi Hong, Haiya Xiang, Juntao Li, Min Zhang, et al.
  ICML 2026, CCF A.
  [Paper] [Code] [Media]
• Revisiting Long-Context Modeling from a Context Denoising Perspective
  Zecheng Tang, Baibei Ji, Juntao Li, Lijun Wu, Haijia Gui, Min Zhang.
  ICLR 2026, CCF A.
  [Paper] [Code] [Media]
• CGMIS: Concept-Graph Based Multi-Hop Instructions Synthesis for Enhancing Long-Context Reasoning
  Zechen Sun, Zecheng Tang, Juntao Li, Wenpeng Hu, Wenliang Chen, Zhunchen Luo, Qiaoming Zhu.
  AAAI 2026, CCF A.
  [Paper]
• L-CiteEval: Do Long-Context Models Truly Leverage Context for Responding?
  Zecheng Tang, Keyan Zhou, Juntao Li, Baibei Ji, Jianye Hou, Min Zhang.
  ACL 2025 (Long Main), CCF A.
  [Paper] [Code] [Media]
• LOGO -- Long cOntext aliGnment via efficient preference Optimization
  Zecheng Tang, Zechen Sun, Juntao Li, Qiaoming Zhu, Min Zhang.
  ICML 2025, CCF A.
  [Paper] [Code] [Media]
• Open-ended long text generation via masked language modeling
  Xiaobo Liang*, Zecheng Tang*, Juntao Li, Min Zhang.
  ACL 2023 (Long Main), CCF A.
  [Paper] [Code]

Generative Model (AIGC)

• Qwen-Image-Layered: Towards Inherent Editability via Layer Decomposition
  Shengming Yin, Zekai Zhang, Zecheng Tang, Kaiyuan Gao, Xiao Xu, Kun Yan, Jiahao Li, Yilei chen, Yuxiang Chen, Heung-Yeung Shum, Lionel Ni, Junyang Lin, Chenfei Wu
  CVPR 2026, CCF A.
  [Paper] [Code] [Media]
• StrokeNUWA: Tokenizing Strokes for Vector Graphic Synthesis
  Zecheng Tang, Chenfei Wu, Zekai Zhang, Mingheng Ni, Juntao Li, Nan Duan, et al.
  ICML 2024, CCF A.
  [Paper] [Code] [Media]
• LayoutNUWA: Revealing the hidden layout expertise of large language models
  Zecheng Tang, Chenfei Wu, Juntao Li, Nan Duan.
  ICLR 2024, CCF A.
  [Paper] [Code] [Media]

Foundation Model

• OpenBA: An Open-sourced 15B Bilingual Asymmetric seq2seq Model
  Juntao Li*, Zecheng Tang*, Yuyang Ding*, Pinzheng Wang*, Min Zhang, et al.
  SCIS 2025, CCF A.
  [Paper] [Code] [Media]
• CMD: a framework for Context-aware Model self-Detoxification
  Zecheng Tang, Keyan Zhou, Juntao Li, Yuyang Ding, Pinzheng Wang, Min Zhang, et al.
  EMNLP 2024 (Long Main), CCF B.
  [Paper] [Code] [Media]
• Improving temporal generalization of pre-trained language models with lexical semantic change
  Zhaochen Su*, Zecheng Tang*, Xinyan Guan, Juntao Li, Lijun Wu, Min Zhang.
  EMNLP 2022 (Long Main), CCF B.
  [Paper] [Code]

---

Selected Preprint

* indicates equal contribution; Full list: Google Scholar.

[Click to hide / view selected paper list]

Technical Report

• Kimi K3: Open Frontier Intelligence
  [Aug 2026] [Arxiv] [Code] [Blog]
• Qwen-Image-2.0-RL Technical Report
  Contributor*
  [Jun 2026] [Arxiv]
• Qwen-Image-2.0 Technical Report
  Core Contributor*
  [May 2026] [Arxiv] [Media]
• Qwen-Image Technical Report
  Core Contributor*
  [Aug 2025] [Arxiv] [Code] [Media]
• Step-Video-T2V Technical Report
  Contributor*
  [Feb 2025] [Arxiv] [Code] [Media]
• Visual chatgpt: Talking, drawing and editing with visual foundation models
  Chenfei Wu, Shengming Yin, Weizhen Qi, Xiaodong Wang, Zecheng Tang, Nan Duan.
  [Mar 2023] [Paper] [Code] [Media]

Long Sequence Modeling (Optimization)

• MemoryRewardBench: Benchmarking Reward Models for Long-Term Memory Management in Large Language Models
  Zecheng Tang, Baibei Ji, Ruoxi Sun, Haitian Wang, Juntao Li, Min Zhang, et al.
  [Jan 2026] [Arxiv] [Code] [Media]
• MMLongCite: A Benchmark for Evaluating Fidelity of Long-Context Vision-Language Models
  Keyan Zhou, Zecheng Tang, Libo Qin, Juntao Li, Min Zhang, et al.
  [Oct 2025] [Arxiv] [Code]
• LongRM: Revealing and Unlocking the Context Boundary of Reward Modeling
  Zecheng Tang, Baibei Ji, Quantong Qiu, Haitian Wang, Xiaobo Liang, Juntao Li, Min Zhang.
  [Oct 2025] [Arxiv] [Code] [Media]
• LOOM-Scope: LOng-cOntext Model evaluation framework
  Zecheng Tang, Haitian Wang, Quantong Qiu, Baibei Ji, Ruoxi Sun, Keyan Zhou, Juntao Li, Min Zhang
  [Jul 2025] [Arxiv] [Code] [Media]

---

Honor & Award

Key terminology translation reference: Chinese Fund Translation.

• 2025: Young Elite Scientists Sponsorship Program (Doctoral Student Special Plan), China Association for Science and Technology (CAST)
• 2025: Top Reviewer, NeurIPS Program Committee.
• 2025: National Scholarship, Soochow University.
• 2024: Star of Tomorrow, Microsoft Research Asia.
• 2022: Outstanding Graduate (ranked 1st/380, GPA 3.95/4.0), Soochow University.
• 2022: Huawei Scholarship (Top 5% Undergraduate), Huawei Inc.

---

Talk

• [Nov 2024] Long Context Modeling in LLMs: Advances and Challenges, NLPCC 2024 Tutorial. [Slides]
• [Apr 2023] Leveraging Large Language Models for Tool Invocation, OPPO (closed-door seminar).

---

Service

I have been serving as a reviewer for the following conferences:

• Natural Language Processing: {ACL, EMNLP, ARR} (2022 - 2026), NLPCC (2023)
• Machine Learning: {ICML, NeurIPS, ICLR} (2024 - 2026)
• Computer Vision: CVPR (2025, 2026), ICCV (2026)
• Artificial Intelligence: AAAI (2022, 2024)

---