🧑‍🔬 Zhiqiang Shen

🏛️ Affiliations

🎓Mohamed bin Zayed University of Artificial IntelligenceAbu Dhabi · United Arab Emirates2021–2026 · 17 paperscurrent🎓Carnegie Mellon UniversityPittsburgh · United States2017–2022 · 28 papers🎓Fudan UniversityShanghai · China2015–2019 · 8 papers🎓Shanghai Jiao Tong UniversityShanghai · China2023 · 4 papers🔬Beijing Institute of Spacecraft Environment EngineeringBeijing · China2019–2023 · 3 papers🎓PLA Army Service AcademyChongqing · China2014–2019 · 3 papers🎓University of Illinois Urbana-ChampaignUrbana · United States2018–2019 · 3 papers🎓National University of Defense TechnologyChangsha · China2002–2025 · 2 papers🎓National University of Science and TechnologyMoscow · Russia2025 · 2 papers🎓National University of Science and TechnologyBulawayo · Zimbabwe2025 · 2 papers

From the bylines of their papers (OpenAlex, 2026-10-14), ORCID and Wikidata; institutions on papers are matched by machine. 11 more institutions on single bylines are not listed.

🧭 Fields, topics and concepts

Of its works in our library, with the number of works on each.

📅 Works by year

2020s

📄 Papers

The most cited first (OpenAlex and our library).

1. SlimPajama-DC: Understanding Data Combinations for LLM Training (2023 · 38 citations)
   Pretraining corpora for LLMs are usually built from several sources, such as web crawls, code, books, and Wikipedia.
2. KITAB-Bench: A Comprehensive Multi-Domain Benchmark for Arabic OCR and Document Understanding (2025 · 21 citations)
   Arabic OCR trails English OCR. The script is cursive and written right to left, and its typography is rich. Existing Arabic datasets mostly cover isolated handwriting or printed lines, not full document understanding. KITAB-Bench fills this gap with 8,809 evaluation samples spread over 9 major tasks and 36…
3. FBI-LLM: Scaling Up Fully Binarized LLMs from Scratch via Autoregressive Distillation (2024 · 7 citations)
   Binarizing LLMs down to pure 1-bit weights could greatly cut memory and compute. Earlier approaches fall short of that limit. Some keep salient weights at higher precision (PB-LLM, BiLLM), and BitNet b1.58 uses ternary values. Others start from a pretrained full-precision model, which locks in its size and vocabulary…
4. LLMSurgeon: Diagnosing Data Mixture of Large Language Models (2026)
   LLM developers rarely disclose the composition of their pretraining data. This makes it hard to audit models for bias, copyright exposure or uneven performance across domains. Membership inference attacks work one document at a time and scale poorly to corpus-level questions. The authors define Data Mixture Surgery…
5. On the Cultural Anachronism and Temporal Reasoning in Vision Language Models (2026)
   Vision-language models are increasingly used in museums, digital archives and educational tools. The authors point to a failure they call cultural anachronism: a model explains a historical object through concepts, materials, techniques or cultural frames from the wrong era. To measure it, they built TAB-VLM, a…

All works

11 works of theirs in our library, the most cited first; 5 with a page here, the others link their original until we summarise them.

1. Partial Is Better Than All: Revisiting Fine-tuning Strategy for Few-shot Learning ↗Proceedings of the AAAI Conference on Artificial Intelligence · no summary yetZhiqiang Shen et al.2021 · 159 citations
2. Vision Transformer Slimming: Multi-Dimension Searching in Continuous Optimization Space ↗arXiv:2201.00814 · no summary yetArnav Chavan et al.2022 · 82 citations
3. SlimPajama-DC: Understanding Data Combinations for LLM TrainingarXiv:2309.10818Zhiqiang Shen, Tianhua Tao, Liqun Ma, et al.2023 · 38 citations
4. KITAB-Bench: A Comprehensive Multi-Domain Benchmark for Arabic OCR and Document UnderstandingFindings of the Association for Computational LinguisticsAhmed Heakl, Muhammad Abdullah Sohail, Mukul Ranjan, et al.2025 · 21 citations
5. Exploring Simple and Transferable Recognition-Aware Image Processing ↗IEEE Transactions on Pattern Analysis and Machine Intelligence · no summary yetZhuang Liu et al.2019 · 14 citations
6. FBI-LLM: Scaling Up Fully Binarized LLMs from Scratch via Autoregressive DistillationarXiv:2407.07093Liqun Ma, Mingjie Sun, Zhiqiang Shen2024 · 7 citations
7. Initializing Models with Larger Ones ↗arXiv (Cornell University) · no summary yetZhiqiu Xu et al.2023 · 7 citations
8. VideoMolmo: Spatio-Temporal Grounding Meets Pointing ↗arXiv (Cornell University) · no summary yetGhazi Shazan Ahmad et al.2025 · 2 citations
9. LLMSurgeon: Diagnosing Data Mixture of Large Language ModelsAnnual Meeting of the Association for Computational Linguistics (ACL)Yaxin Luo, Jiacheng Cui, Xiaohan Zhao, et al.2026
10. On the Cultural Anachronism and Temporal Reasoning in Vision Language ModelsFindings of the Association for Computational LinguisticsMukul Ranjan, Prince Jha, Khushboo Kumari, et al.2026
11. Mobile-MMLU: A Mobile Intelligence Language Understanding Benchmark ↗arXiv (Cornell University) · no summary yetSondos Mahmoud Bsharat et al.2025

🤝 Co-authors here

12 of 14 co-authors with a page, the most papers together first.

• Liqun Ma ×2
• Mukul Ranjan ×2
• Ahmed Heakl
• Bowen Tan
• Eric P. Xing
• Fahad Shahbaz Khan
• Jiacheng Liu
• Joel Hestness
• Mingjie Sun
• Natalia Vassilieva
• Rania Elbadry
• Salman Khan

Venues

Where their works here were published, the most works first, with the editions of those that have a page.

• Findings of the Association for Computational Linguistics ×2: Findings of the Association for Computational Linguistics 2026, Findings of the Association for Computational Linguistics 2025
• Annual Meeting of the Association for Computational Linguistics (ACL): ACL 2026
• IEEE Transactions on Pattern Analysis and Machine Intelligence
• Proceedings of the AAAI Conference on Artificial Intelligence

🔗 Cited by

Papers of our library cite their works 79 times (of the 149,439 papers whose bibliographies we read); 24 citing papers with a page here →.

Updated 2 November 2026

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/authors/zhiqiang-shen/