PeopleofSuperintelligence

Researcher

Min-Hung Chen

Also known as Steve Chen, 陳敏弘

Min-Hung (Steve) Chen is a multimodal AI researcher whose work spans video understanding, domain adaptation and efficient fine-tuning. After earning degrees at National Taiwan University in 2010 and 2012, he took a PhD in electrical and computer engineering at Georgia Tech in 2020 under Ghassan AlRegib, with video domain adaptation papers at ICCV 2019 and CVPR 2020. He worked at MediaTek and Microsoft Azure AI before joining NVIDIA Research Taiwan in 2022, where he coauthored the DoRA fine-tuning method, an ICML 2024 oral paper, and the Hymba small language model, an ICLR 2025 paper. A 2019 paper from his PhD years, TS-LSTM and temporal-inception, received a 2023 EURASIP Best Paper Award. Since March 2026, he is a Staff Research Scientist at NVIDIA Research Taiwan, working on Vision+X multimodal AI.

Now at
NVIDIA
Born
1944
Works
61
Citations
1,090
h-index
15
Badges

Years link to their sources. Publication numbers from OpenAlex, as of 22 October 2026.

Timeline

  1. 2010–2014

    Degrees at National Taiwan University

    He earns a BSc in electrical engineering at National Taiwan University in 2010 and an MSc in integrated circuits and systems there in 2012. From July 2013 to July 2014, he is a research assistant at Academia Sinica in Taipei, working on multi-modal action recognition.

  2. 2014–2020

    A Georgia Tech doctorate

    From August 2014, he studies at Georgia Tech, where his research covers video understanding beyond full supervision, human action understanding and robust machine learning for autonomous vehicles. Ghassan AlRegib advises the work, in collaboration with Zsolt Kira, and internships at PlayStation in 2018 and Baidu USA in 2019 address cross-domain action recognition and segmentation. He completes his PhD in electrical and computer engineering in 2020.

  3. 2019–2020

    Video domain adaptation papers

    In 2019, he is first author of Temporal Attentive Alignment for Large-Scale Video Domain Adaptation, an ICCV 2019 oral paper that introduces the UCF-HMDB_full and Kinetics-Gameplay datasets and the TA3N network. His 2020 CVPR paper, Action Segmentation with Joint Self-Supervised Temporal Domain Adaptation, proposes SSTDA and outperforms the prior state of the art on the GTEA, 50Salads and Breakfast benchmarks.

  4. 2020–2022

    MediaTek and Microsoft Azure AI

    In October 2020, he joins MediaTek in Hsinchu as a Senior AI Engineer, researching methods for Edge-AI and co-organizing the Learned Smartphone ISP Challenge at the CVPR 2021 Mobile AI Workshop. In January 2022, he joins the Face Science Team at Microsoft Azure AI as a Research Engineer II, working on generalizable and explainable facial liveness approaches for cloud services.

  5. 2022

    Joining NVIDIA Research Taiwan

    NVIDIA Research Taiwan, founded in 2022, is the APAC research team of NVIDIA Research and is led by Professor Frank Wang. He joins it in 2022 as a Senior Research Scientist working on Vision+X multimodal AI.

  6. 2023

    EURASIP Best Paper Award

    The paper TS-LSTM and temporal-inception, which he coauthors with Chih-Yao Ma, Zsolt Kira and Ghassan AlRegib and publishes in Signal Processing: Image Communication in 2019, receives a 2023 EURASIP Best Paper Award. It presents two methods, TS-LSTM and Temporal-Inception, that exploit spatiotemporal dynamics for activity recognition.

  7. 2024–2025

    DoRA and Hymba

    In 2024, he coauthors DoRA, a parameter-efficient fine-tuning method that splits a pretrained weight into magnitude and direction and applies LoRA to the directional updates; it outperforms LoRA on LLaMA, LLaVA and VL-BART tasks and is accepted to ICML 2024 as an oral paper. He introduces DoRA in an NVIDIA Technical Blog post in June 2024. In 2025, he coauthors Hymba, an ICLR 2025 family of small language models whose hybrid-head architecture integrates attention mechanisms with state space models.

  8. Since March 2026

    Staff Research Scientist at NVIDIA

    He is promoted to Staff Research Scientist at NVIDIA Research Taiwan in March 2026, where his work on Vision+X multimodal AI covers vision-language, 4D and spatial understanding, vision-language-action models and agentic AI. He also serves as an Area Chair for ICML 2026, ECCV 2026 and NeurIPS 2026.

Career

  1. Since 2022NVIDIASenior Research Scientist, Staff Research Scientist
  2. 2022MicrosoftResearch Engineer II
  3. 2020–2022MediaTekSenior AI Engineer
  4. 2019Baidu USAIntern
  5. 2018PlayStationIntern
  6. 2014–2022National Taiwan University
  7. 2013–2014Academia SinicaResearch Assistant

Education

  1. 2020Georgia Institute of TechnologyPhD
  2. 2012National Taiwan UniversityBSc and MSc

From the biography and OpenAlex affiliations (the years of papers that list the organization), so some years are approximate.

Most-cited works

  1. TS-LSTM and temporal-inception: Exploiting spatiotemporal dynamics for activity recognition2018 · Signal Processing Image Communication · 247 citations
  2. Temporal Attentive Alignment for Large-Scale Video Domain Adaptation2019 · IEEE/CVF International Conference on Computer Vision (ICCV) · 178 citations
  3. Action Segmentation With Joint Self-Supervised Temporal Domain Adaptation2020 · IEEE Computer Society Conference on Computer Vision and Pattern Recognition · 123 citations
  4. Traffic Sign Detection Under Challenging Conditions: A Deeper Look into Performance Variations and Spectral Characteristics2019 · IEEE Transactions on Intelligent Transportation Systems · 88 citations
  5. Holistic Interaction Transformer Network for Action Detection2023 · IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) · 52 citations

From OpenAlex, as of 21 October 2026.

  1. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes Min-Hung Chen. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other university. We compare doctorate and 5+ year employment links, career timing and current workplaces for 187 major-prize and 1,111 best-paper winners.
  2. Google has the highest combined h-index among AI organizationsIncludes NVIDIA. We estimate that Google holds the highest combined h-index at 8,016, while universities account for 73% of the deduplicated total across AI organizations.
  3. 20% of people at companies who shape AI also hold a university role, led by Google and Stanford UniversityIncludes NVIDIA. We estimate that 539 of the 6,766 people with a known current organization (8%) hold both a university role and a company role, including 20% of people at companies and 13% of people at universities.
  4. Google alumni have founded or led 133 other organizations, more than any other AI labIncludes NVIDIA. We estimate that alumni of 17 major AI labs went on to found or lead 350 other organizations, including 226 they founded, and we trace 472 career moves directly between 23 labs since 2010.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Min-Hung Chen yet. Send us a free-licensed photo.

  1. Min-Hung Chen minhungchen.netlify.app. Checked 20 October 2026.
  2. 【2025-12-03】Dr. Min-Hung Chen 陳敏弘 / NVIDIA / Multimodal Efficient AI Research at NVIDIA Taiwan csie.ntu.edu.tw. Checked 20 October 2026.
  3. Temporal Attentive Alignment for Large-Scale Video Domain Adaptation arxiv.org. Checked 20 October 2026.
  4. Action Segmentation with Joint Self-Supervised Temporal Domain Adaptation arxiv.org. Checked 20 October 2026.
  5. NVIDIA Research Taiwan research.nvidia.com. Checked 20 October 2026.
  6. EURASIP Newsletter – Vol. 34, April 2023 eurasip.org. Checked 20 October 2026.
  7. DoRA: Weight-Decomposed Low-Rank Adaptation | NVIDIA Research Taiwan research.nvidia.com. Checked 20 October 2026.
  8. Introducing DoRA, a High-Performing Alternative to LoRA for Fine-Tuning | NVIDIA Technical Blog developer.nvidia.com. Checked 20 October 2026.
  9. Hymba: A Hybrid-head Architecture for Small Language Models | NVIDIA Research Taiwan research.nvidia.com. Checked 20 October 2026.
  10. Min-Hung Chen | NVIDIA Research Taiwan research.nvidia.com. Checked 20 October 2026.