Bolei Zhou
Associate Professor
Directs the Vision and Autonomy Intelligence Lab on interpretable computer vision, generative models and simulation for robot learning and driving.
Bolei Zhou's lab at UCLA works on embodied & active vision, generative & foundation models and trustworthy & safe ML. 62 papers since 2023, 47 at top venues, mostly CVPR and NeurIPS. 9 PhD students. 16 alumni, most often to Alibaba and Amazon. Applications close .
At a glance
- papers, 2023–2026
- 16/ year
- at T1 venues
- 76%
- PhD students
- 9
- MS students
- 0
- alumni listed
- 16
Papers by yearfrom the lab homepage, 2023–2026
- 16 papers2026 · T1 12 · collecting
- 19 papers2025 · T1 18
- 15 papers2024 · T1 7
- 12 papers2023 · T1 10
T1 venuesOtherVenues: CVPR, NeurIPS, CoRL, ICCV, ICLR
Research focuspaper titles that use each topic's words
- Generative & foundation models8 papers
- Embodied & active vision6 papers
- autonomy3 papers
- interpretable2 papers
- Trustworthy & safe ML1 paper
Paper list
From the lab homepage.
202616 papers
- CVPR (T1)AnimaMimic: Imitating 3D Animation from Video Priors
- CVPR (T1)AURA: Multi-modal Shared Autonomy for Urban Navigation
- CoRL (T1)BridgeSim: Unveiling the OL-CL Gap in End-to-End Autonomous Driving
- CoRL (T1)Cue the Flow: Steering Flow-Matching Policies for Open-World Delivery Manipulation
- HRI (T1)DiSCo: Diffusion Sequence Copilots for Shared Autonomy
- CoRL (T1)DreamStream: Towards Policy-Oriented Generative Simulation for End-to-End Driving
- CoRL (T1)From Imitation to Alignment: Human-Preference Flow Policies for Long-Horizon Sidewalk Navigation
- International Conference on Learning RepresentationFrom Seeing to Experiencing: Scaling Navigation Foundation Models with Reinforcement Learning
- CVPR (T1)Group Diffusion: Enhancing Image Generation by Unlocking Cross-Sample Collaboration
- International Conference on Learning RepresentationJoint Optimization for 4D Human-Scene Reconstruction in the Wild
- ICRA (T1)Learning Sidewalk Autopilot from Multi-Scale Imitation with Corrective Behavior Expansion
- ECCV (T1)QuantV2X: A Fully Quantized Multi-Agent System for Cooperative Perception
202519 papers
- ICLR (T1)3DitScene: Editing Any Scene via Language-guided Disentangled Gaussian Splatting
- NeurIPS (T1)Adv-BMT: Bidirectional Motion Transformer for Safety-Critical Traffic Scenario Generation
- ICML (T1)AIM: Robot-Gated Interactive Imitation Learning with Adaptive Intervention Mechanism
- NeurIPS (T1)AutoVLA: A Vision-Language-Action Model for End-to-End Autonomous Driving with Adaptive Reasoning and Reinforcement Fine-Tuning
- RA-LBEVCon: Advancing Bird’s Eye View Perception With Contrastive Learning
- IROS (T1)CooPre: Cooperative Pretraining for V2X Cooperative Perception
- ICRA (T1)Data-Efficient Learning from Human Interventions for Mobile Robots
- CVPR (T1)Embodied Scene Understanding for Vision Language Models via MetaVQA
- ICLR (T1)Learning to Generate Diverse Pedestrian Movements from Web Videos with Noisy Labels
- ICLR (T1)MetaUrban: An Embodied AI Simulation Platform for Urban Micromobility
- ICCV (T1)Occupancy Learning with Spatiotemporal Memory
- NeurIPS (T1)Predictive Preference Learning from Human Interventions
202415 papers
- Nature Biomedical EngineeringAccurate prediction of disease-risk factors from volumetric medical scans by a deep vision model pre-trained with 2D scans
- IEEE/CVF Computer Vision and Pattern Recognition ConferenceBerfscene: Bev-conditioned equivariant radiance fields for infinite 3d scene generation
- RA-LBEVGen: Street-View Image Generation from a Bird’s-Eye View Layout
- NeurIPS (T1)Ctrl-X: Controlling Structure and Appearance for Text-To-Image Generation Without Guidance
- 3DVEfficient 3d articulated human generation with layered surface volumes
- Nature (T1)Experiment-free Exoskeleton Assistance via Learning in Simulation
- IEEE/CVF Computer Vision and Pattern Recognition ConferenceFreeControl: Training-Free Spatial Control of Any Text-to-Image Diffusion Model with Any Condition
- TPAMI (T1)In-Domain GAN Inversion for Faithful Reconstruction and Editability
- Communications EngineeringLocalization and recognition of human action in 3D using transformers
- IEEE/CVF Computer Vision and Pattern Recognition ConferenceScenewiz3d: Towards text-guided 3d scene composition
- NeurIPS (T1)Shared Autonomy with IDA: Interventional Diffusion Assistance
- NeurIPS (T1)SimGen: Simulator-conditioned Driving Scene Generation
202312 papers
- CoRL (T1)CAT: Closed-loop Adversarial Training for Safe End-to-End Driving
- TMLR (T1)ChemSpacE: Interpretable and Interactive Chemical Space Exploration
- CVPR (T1)DisCoScene: Spatially Disentangled Generative Radiance Fields for Controllable 3D-aware Scene Synthesis
- TPAMI (T1)GH-Feat: Learning Versatile Generative Hierarchical Features From GANs
- International Conference on Learning RepresentationGuarded Policy Optimization with Imperfect Online Demonstrations
- NeurIPS (T1)Learning from Active Human Involvement through Proxy Value Propagation
- ICCV (T1)One-shot generative domain adaptation
- Thirty- Conference on Neural Information Processing Systems Datasets and Benchmarks TrackScenarioNet: Open-Source Platform for Large-Scale Traffic Scenario Simulation and Modeling
- ICLR (T1)Towards Smooth Video Composition
- ICRA (T1)TrafficGen: Learning to Generate Diverse and Realistic Traffic Scenarios
- CVPR (T1)V2V4Real: A large-scale real-world dataset for Vehicle-to-Vehicle Cooperative Perception
- ICRA (T1)V2XP-ASG: Generating Adversarial Scenes for Vehicle-to-Everything Perception
Where alumni went16 listed, 44% to industry
Industry 4Academia 5
Most often to
- Alibaba
- Amazon
Also to
- UCLA
- HKUST
- Stanford University
Apply to UCLAFall 2027
Computer Science PhD
- Round
- Single round (Dec 15)
- Opens
- —
- GRE
- optional
- English
- See program page
- Fee
- See program page
Not yet confirmed for this intake.
Similar labsnearest by research summary
- CNPeking
Hao Tang
- Vision-language & multimodal
- DETUM
Daniel Cremers
- 3D vision & neural rendering
- USCMU
Anthony Rowe
- Image & video generation
- CNTsinghua
Jiwen Lu
- Embodied & active vision
- CNZhejiang
Min Zhang
- Rendering & appearance
- CNZhejiang
Weidong Geng
- Physics-based simulation & animation
Your notes
Save this lab to keep notes.
Data
Checked against the lab homepage.
- Collected
- Verified
- Papers
- The lab homepage
- Members
- Lab homepage
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.
UCLAcloses