Profile

Hey there. I am Di Chen, an algorithm researcher who has worked at Alibaba Tongyi Vision Intelligence Lab and Agibot Research.

I got my Ph.D. degree in Computer Science at Nanjing University of Science and Technology. My supervisors are Prof. Jian Yang and Prof. Shanshan Zhang. During 2019 – 2020, I was a visiting scholar at Max-Planck-Institut für Informatik, supervised by Prof. Dr. Bernt Schiele.

My research interests encompass computer vision and embodied AI, spanning person search (a task that combines pedestrian detection and person re-identification), multiple object tracking, face recognition, and text-to-image/video generation. At present, I’m focusing on world models and generative simulators for robotic manipulation.

Experience

Algorithm Researcher Agibot | 2025.05 – 2026.04 • Video Generation • World Models • Generative Simulator • Robotic Manipulation

Algorithm Researcher Alibaba | 2021.07 – 2025.04 • Tongyi Wanxiang (通义万相): A comprehensive image and video generation service • Diffusion model distillation & acceleration • Neural Radiance Fields, 3D reconstruction • Large-scale ****diffusion models for Taobao image generation and inpainting/outpainting • Multi-modal retrieval for large scale data collection/filtering

Publications

World Models & Robotic Manipulation

τ₀-WM: A Unified Video-Action World Model for Robotic Manipulation, 2026, arXiv bibtex pdf
GE-Sim 2.0: A Roadmap Towards Comprehensive Closed-loop Video World Simulators for Robotic Manipulation, 2026, arXiv bibtex code page pdf
Real2Edit2Real: Generating Robotic Demonstrations via a 3D Control Interface, 2026, CVPR bibtex page pdf
Act2Goal: From World Model To General Goal-conditioned Policy, 2026, RSS bibtex page pdf

Visual Generation

Wan: Open and advanced large-scale video generative models, 2025, arXiv bibtex code page pdf
Ranni: Taming Text-to-Image Diffusion for Accurate Instruction Following, 2024, CVPR (Oral) bibtex code page pdf
Composer: Creative and Controllable Image Synthesis with Composable Conditions, 2023, ICML page pdf

Neural Radiance Fields

GeoAug: Data Augmentation for Few-Shot NeRF with Geometry Constraints, 2022, ECCV bibtex pdf poster

Person Search/Detection/Re-Identification

FSLoRA: Harmonizing Detection and Re-Identification via Freq-Spatial Low-Rank Adapter for One-Stage Person Search, 2026, CVPR bibtex pdf
Divide and Conquer: Hybrid Pre-training for Person Search, 2024, AAAI bibtex code pdf
Keypoint Message Passing for Video-based Person Re-Identification, 2022, AAAI (Oral) bibtex pdf
PoseTrack21: A Dataset for Person Search, Multi-Object Tracking and Multi-Person Pose Tracking, 2022, CVPR bibtex pdf
Grouped Adaptive Loss Weighting for Person Search, 2022, ACM MM bibtex pdf
Norm-Aware Embedding for Efficient Person Search and Tracking, 2021, IJCV bibtex pdf
Guided Attention in CNNs for Occluded Pedestrian Detection and Re-Identification, 2021, IJCV bibtex pdf
Norm-Aware Embedding for Efficient Person Search, 2020, CVPR bibtex code pdf poster
Hierarchical Online Instance Matching for Person Search, 2020, AAAI bibtex code pdf poster slides
Person Search by Separated Modeling and A Mask-Guided Two-Stream CNN Model, 2020, TIP bibtex pdf
Person Search via A Mask-Guided Two-Stream CNN Model, 2018, ECCV bibtex blog pdf poster

Representation Learning

Animating Images to Transfer CLIP for Video-Text Retrieval, 2022, SIGIR bibtex pdf
Joint Bayesian Guided Metric Learning for End-to-end Face Verification, 2018, Neurocomputing bibtex pdf