Na Zhao

Assistant Professor@Singapore University of Technology and Design·新加坡·OpenReview

研究方向

spatial intelligence · 3D generation · 3D reconstruction · continual learning · few-shot learning · semi supervised learning · 3D object detection · 3D scene understanding · 3D semantic segmentation · computer vision

4.7

TokenSwap: Backdoor Attack on the Compositional Understanding of Large Vision-Language Models

ICLR 2026Rejected

4.0

Scene-R1: Video-Grounded Large Language Models for 3D Scene Reasoning without 3D Annotations

ICLR 2026Withdrawn

7.3

AffordBot: 3D Fine-grained Embodied Reasoning via Multimodal Large Language Models

NeurIPS 2025Poster

通讯

6.1

How Do Images Align and Complement LiDAR? Towards a Harmonized Multi-modal 3D Panoptic Segmentation

ICML 2025Poster

通讯

5.6

GaussianBlock: Building Part-Aware Compositional and Editable 3D Scene by Primitives and Gaussians

ICLR 2025Poster

通讯

4.7

Never Forget the Basics: In-distribution Knowledge Retention for Continual Test-time Adaptation in Human Motion Prediction

合作者 (20)