Kaiming He

Kaiming He is a Chinese computer scientist and researcher renowned for contributions to deep learning and computer vision, particularly in convolutional neural network architectures and object detection. He has worked at leading research labs and influenced modern image recognition, segmentation, and representation learning through widely cited papers and open-source code.

Early life and education

He was born in Jiujiang and educated in China. He earned undergraduate and graduate degrees from Tsinghua University and pursued doctoral studies associated with the Chinese Academy of Sciences and the University of Chinese Academy of Sciences, conducting research intersecting with groups at Microsoft Research Asia and collaborating with researchers from Hong Kong University of Science and Technology and the Chinese University of Hong Kong. During his formative years he interacted with scholars linked to institutions such as Peking University, Zhejiang University, Shanghai Jiao Tong University, and research centers like Baidu Research and Tencent AI Lab.

Research career and positions

He has held positions at major industrial and academic organizations including Microsoft Research, Facebook AI Research, SenseTime, Amazon labs, and affiliations with universities such as Stanford University, Massachusetts Institute of Technology, University of Oxford, and ETH Zurich through collaborations and visiting appointments. His work sits at the intersection of projects from groups like ImageNet teams, COCO organizers, and partnerships with companies such as Google Research, NVIDIA, and Intel Labs. He has participated in conferences and workshops sponsored by NeurIPS, ICCV, CVPR, ECCV, ICLR, and institutions such as IEEE and ACM.

Key contributions and major works

He is best known for inventing or co-inventing influential architectures and methods including residual networks used in conjunction with datasets like ImageNet and benchmarks such as MS COCO and PASCAL VOC. He co-authored landmark systems including ResNet models, Mask R-CNN, and advances in normalization techniques linked to work on Batch Normalization and representation learning. His contributions informed frameworks and libraries such as Caffe, PyTorch, TensorFlow, and evaluation toolkits derived from OpenCV and implementations running on hardware by NVIDIA and Intel accelerators. Collaborators include researchers associated with Ross Girshick, Shaoqing Ren, Jian Sun, Piotr Dollár, and groups from Carnegie Mellon University, University of California, Berkeley, and Google Brain.

Awards and honors

He has received recognition from professional societies and conferences, including distinctions from IEEE and ACM, paper awards at venues like CVPR and ICCV, and honors tied to impactful contributions to datasets such as ImageNet and MS COCO. He has been invited to keynote at symposia hosted by NeurIPS, ICML, and AAAI, and has been listed among notable researchers in rankings maintained by entities like Google Scholar, Microsoft Academic, and academic consortia associated with Clarivate Analytics.

Selected publications and impact

Selected influential publications include papers presenting residual learning for deep networks (ResNet), region-based convolutional approaches culminating in Mask R-CNN, and studies on deep residual learning, network initialization, and representation transfer evaluated on benchmarks like ImageNet Large Scale Visual Recognition Challenge, COCO Challenge, and PASCAL Visual Object Classes Challenge. These works have driven progress in applications across companies and labs such as Facebook, Microsoft, Google, Amazon Web Services, Tesla, Waymo, and research groups at Uber AI Labs and DeepMind. His papers are highly cited on platforms like Google Scholar and indexed by bibliographic services including DBLP and Semantic Scholar, and have influenced tooling in ecosystems maintained by GitHub and ArXiv.

Category:Computer scientists Category:Chinese computer scientists