Ming-Hsuan YangResearch Scientist · Google DeepMindResearcher↗
Researcher · Academic
MITAssociate Professor
Google DeepMindDistinguished Scientist
Kaiming He co-created ResNet, whose shortcut connections made much deeper neural networks practical to train.[1] He also helped develop Mask R-CNN, which locates and outlines individual objects,[2] and masked autoencoders, which teach visual AI by reconstructing missing image patches rather than relying on human labels.[3]
MITFeb 2024-present
Google DeepMindJun 2025-present
Meta2016-2024Work and education history is primarily focused on AI-relevant roles and may not be comprehensive.
ResNet, an image-recognition network, uses shortcuts between layers to make much deeper networks easier to train.
Shared neural-network processing made locating and identifying objects faster with Faster R-CNN.
Focal loss shifts object-detector training toward difficult examples rather than easy ones.
First authorship; residual framework enables substantially deeper networks.
First authorship; self-supervised reconstruction of masked patches.
He is first author of MoCo; contrastive image representations trained without labels outperform supervised pretraining on seven detection/segmentation tasks in the paper.
Based on public sources. If accuracy is important, verify details using the links or references.
For corrections, updates, or requests to remove information, contact us.
Reviewed August 7, 2026 · Checked for accuracy and new information.