Scholar
Ju He
Google Scholar ID: NyTPm_zUV_kC
Applied Scientist, Amazon Frontier AI & Robotics (FAR)
Citations & Impact
Citations
1,668
H-index
15
i10-index
19
Publications
20
Co-authors
28
list available
Publications
Resume (English only)
Academic Achievements
- - Paper 'FlowTok: Flowing Seamlessly Across Text and Image Tokens' accepted to ICCV 2025
- - Paper 'Democratizing Text-to-Image Masked Generative Models with Compact Text-Aware One-Dimensional Tokens' accepted to ICCV 2025
- - Paper 'Randomized Autoregressive Visual Generation' accepted to ICCV 2025
- - Paper 'Beyond Next-Token: Next-X Prediction for Autoregressive Visual Generation' accepted to ICCV 2025
- - Paper 'Alleviating Distortion in Image Generation via Multi-Resolution Diffusion Models' accepted to NeurIPS 2024
- - Paper 'A Simple Video Segmenter by Tracking Objects Along Axial Trajectories' published in TMLR 2024
- - Paper 'Convolutions Die Hard: Open-Vocabulary Segmentation with Single Frozen Convolutional CLIP' accepted to NeurIPS 2023
- - Paper 'Compositor: Bottom-up Clustering and Compositing for Robust Part and Object Segmentation' accepted to CVPR 2023
- - Paper 'CORL: Compositional Representation Learning for Few-Shot Classification' accepted to WACV 2023
- - Paper 'PartImageNet: A Large, High-Quality Dataset of Parts' accepted to ECCV 2022
- - Paper 'TransFG: A Transformer Architecture for Fine-grained Recognition' accepted to AAAI 2022
- - Paper 'Semi-synthesis: A fast way to produce effective datasets for stereo matching'
Research Experience
- - Applied Scientist at Amazon Frontier AI & Robotics (FAR)
- - Research Scientist at ByteDance Seed
Education
- Received Ph.D. in Computer Science from Johns Hopkins University, advised by Bloomberg Distinguished Professor Alan L. Yuille; B.S. in Computer Science from Peking University.
Background
- Currently an Applied Scientist at Amazon Frontier AI & Robotics (FAR). Previously, worked as a Research Scientist at ByteDance Seed.
