PeopleofSuperintelligence

Researcher

Bin Xiao

Also known as Xiao Bin, leoxiaobin

Bin Xiao is a computer vision researcher whose work includes human-pose estimation, vision foundation models and agentic AI. His early papers cover Simple Baselines in 2018, HRNet in 2019, HigherHRNet in 2020 and CvT in 2021. He coauthored Florence in November 2021 and led Florence-1 and Florence-2 model work from 2020 to 2023. In 2024, he coauthored Florence-2 and the Phi-3 Technical Report while leading Phi-3 Vision and Phi-3.5 Vision. Since 2025, he has worked on post-training for reasoning, coding and agentic models; as of October 2026, he is a Member of Technical Staff on Microsoft's Superintelligence Team in Redmond.

Now at
Microsoft

Timeline

  1. 2018

    Simple Baselines pose methods

    At ECCV in 2018, he coauthors Simple Baselines for Human Pose Estimation and Tracking, which provides simple baseline methods for pose estimation and tracking.

  2. 2019

    HRNet pose estimation

    At CVPR in 2019, he coauthors Deep High-Resolution Representation Learning for Human Pose Estimation. HRNet maintains high-resolution representations throughout the network through parallel multi-resolution branches and repeated feature fusion.

  3. 2020

    HigherHRNet pose estimation

    At CVPR in 2020, he coauthors HigherHRNet, a bottom-up pose-estimation method that learns scale-aware representations using high-resolution feature pyramids.

  4. 2021

    CvT vision transformer

    At ICCV in 2021, he coauthors CvT, a Vision Transformer architecture that introduces convolutions to improve performance and efficiency.

  5. November 2021

    Florence foundation models

    In November 2021, he coauthors Florence, a computer-vision foundation model for classification, retrieval, object detection, VQA, image captioning, video retrieval and action recognition. From 2020 to 2023, he leads Florence-1 and Florence-2 model work. His research interests include unified vision-language representations for captioning, OCR, grounding and document understanding.

  6. 2024

    Florence-2 and Phi-3 Vision

    At CVPR in June 2024, he coauthors Florence-2, a prompt-based vision foundation model for captioning, object detection, grounding and segmentation. In 2024, he leads Phi-3 Vision and Phi-3.5 Vision and coauthors the Phi-3 Technical Report.

  7. Since 2025

    Post-training agentic models

    He begins post-training work for reasoning, coding and agentic models in 2025, including planning and tool use.

  8. As of October 2026

    Microsoft Superintelligence Team

    As of October 2026, he is a Member of Technical Staff on Microsoft's Superintelligence Team in Redmond. He researches multimodal foundation models, computer vision and agentic AI.

  1. Among employers with 30+ full-time cases, OpenAI leads primary technical role H-1B pay at $310,000, and Amazon, Google and Microsoft file 57% of all casesIncludes Microsoft. Across 440,669 certified US H-1B filings from 44 employers (FY2020–FY2026), OpenAI leads primary technical role base pay at $310,000 and xAI leads overall base pay at $290,000.
  2. Google had the most people shaping AI every year from 2014 to 2025Includes Microsoft. Our estimates of which organizations employed the most people shaping AI each year from 1955 to 2025, tracing the shift from university labs to companies.
  3. The company share of AI researchers’ top papers was 31% in 2020–2024 and 8% in 1970–1974Includes Microsoft. We estimate where the people shaping AI worked when their most-cited papers appeared: the company share was 31% in 2020–2024 and 8% in 1970–1974, and Google led all organizations in 2020–2024.
  4. Google has the highest combined h-index among AI organizationsIncludes Microsoft. 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.

Sources and credits

Photo: We have no published photo of Bin Xiao yet. Send us a free-licensed photo.

  1. Simple Baselines for Human Pose Estimation and Tracking openaccess.thecvf.com. Checked 22 October 2026.
  2. Deep High-Resolution Representation Learning for Human Pose Estimation openaccess.thecvf.com. Checked 22 October 2026.
  3. HigherHRNet: Scale-Aware Representation Learning for Bottom-Up Human Pose Estimation openaccess.thecvf.com. Checked 22 October 2026.
  4. CvT: Introducing Convolutions to Vision Transformers openaccess.thecvf.com. Checked 22 October 2026.
  5. Bin Xiao — Research & Publications leoxiaobin.github.io. Checked 22 October 2026.
  6. Florence: A New Foundation Model for Computer Vision microsoft.com. Checked 22 October 2026.
  7. Florence-2: Advancing a Unified Representation for a Variety of Vision Tasks microsoft.com. Checked 22 October 2026.
  8. [2404.14219] Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phone arxiv.org. Checked 22 October 2026.