Scholar
Liang-Chieh Chen
Google Scholar ID: ACjYGPUAAAAJ
Senior Principal Scientist, Amazon FAR, LA
Citations & Impact
Citations
115,073
H-index
42
i10-index
63
Publications
20
Co-authors
129
list available
Contact
Publications
Resume (English only)
Academic Achievements
- Led or contributed to influential DeepLab derivatives: Auto-DeepLab, Panoptic-DeepLab (winner of ICCV 2019 Mapillary Vistas Panoptic Segmentation track and top performer on Cityscapes), Axial-DeepLab, ViP-DeepLab, MaX-DeepLab, kMaX-DeepLab.
- Co-developed MobileNetv2 and MobileNetv3, now standard architectures for efficient mobile vision models.
- Multiple papers accepted at top venues including ICCV 2025 and NeurIPS 2025, covering diffusion transformers, physics-based video generation, panoptic captioning, flow-based generative models, etc.
- Served as Area Chair for ICCV (2019, 2025), CVPR (2020, 2023–2026), ECCV (2020, 2024), NeurIPS (2022, 2024–2025), ICML (2025).
- Action Editor for Transactions on Machine Learning Research (TMLR).
- Received Outstanding Reviewer awards at CVPR 2021 and 2018, and Top Reviewer at NeurIPS 2019.
Background
- Currently a Research Scientist at Apple AI/ML, building cutting-edge visual generative models.
- Widely recognized for co-developing the DeepLab series (DeepLabv1 to DeepLabv3+) with George Papandreou.
- Introduced atrous (dilated) convolution in December 2014, now a foundational technique for dense prediction tasks.
- Known for collaborative work on MobileNetv2 and MobileNetv3, which have become standards for efficient mobile neural networks.
- His team focuses on fundamental research in computer vision, visual generation, and representation learning.
