Elamur Research⌕
☰
home / authors / Mingxuan Yuan

Mingxuan Yuan

Mingxuan Yuan has 3 papers in our library, in 2026, most of them on Efficient models & compression; often with Hui-Ling Zhen, Xianzhi Yu, Pengxiang Zhao.

3 papers in our library · 2026

OpenAlex · ORCID · Wikidata · h-index 26 · 191 works · 2,270 citations OpenAlex, 2026-10-14

📄 3 works in our library · 🔗 Cited 1 time in our library · ⭐ Paper Stars ranking: #11077

The profile is the library's match of the name; two people with one name are not always told apart.

Elsewhere: Google Scholar · Semantic Scholar · dblp · ACL Anthology

 Affiliations

🏢Huawei TechnologiesShenzhen · China2012–2026 · 167 paperscurrent🎓Hong Kong University of Science and TechnologyHong Kong · China2007–2016 · 23 papers🏢Huawei TechnologiesStockholm · Sweden2024–2026 · 7 papers🎓City University of Hong KongHong Kong · China2023–2026 · 4 papers🎓Nanjing UniversityNanjing · China2025 · 4 papers🎓Chinese University of Hong KongHong Kong · China2023–2024 · 2 papers
🏛️Huawei Noah's Ark LabHong Kong · ChinaEmployment

From the bylines of their papers (OpenAlex, 2026-10-14), ORCID and Wikidata; institutions on papers are matched by machine. 6 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.

📚 Shelves 8The field of science this work belongs to (OpenAlex fields).Computer Science · 3 worksComputer Vision and Pattern Recognition · 2 worksArts and Humanities · 2 worksEngineering · 1 workArtificial Intelligence · 2 worksLanguage and Linguistics · 2 worksComputational Theory and Mathematics · 1 workElectrical and Electronic Engineering · 1 work

🧭 Topics 1Research subjects and languages the work is about.Efficient models & compression · 1 work

 Works by year

2020s

20263 papers

 Papers

The most cited first (OpenAlex and our library).

  1. Analytical FFN-to-MoE Restructuring via Activation Pattern Analysis (2026 · 1 citation)
    Most inference compute in large language models goes to the feed-forward (FFN) blocks. Converting a dense model into a Mixture-of-Experts (MoE) can cut that cost, but earlier conversion methods needed continual training on billions of tokens. Pei et al. instead restructure FFNs after training, in a single analytical…
  2. Unleashing Low-Bit Inference on Ascend NPUs: A Comprehensive Evaluation of HiFloat Formats (2026 · 1 citation)
    The paper asks whether Huawei's HiFloat number formats, built for Ascend NPUs, work well for post-training quantization (PTQ) of large language models.
  3. RIFT: Repurposing Negative Samples via Reward-Informed Fine-Tuning (2026)
    Post-training LLMs with Supervised Fine-Tuning requires expensive expert demonstrations. Rejection Sampling Fine-Tuning (RFT) avoids this by training on the model's own verified outputs, but it throws away every incorrect sample, which wastes compute and ignores information about failure modes. The authors propose…

 Co-authors here

6 co-authors with a page, the most papers together first.

Venues

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

 Cited by

Papers of our library cite their works 1 time (of the 149,439 papers whose bibliographies we read).

Updated

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