Yan Li

Yan Li (李言)

Research Fellow, National University of Singapore · Incoming Assistant Professor, Hong Kong University of Science and Technology (Guangzhou)

My research focuses on robotics, spatial AI, SLAM, 3D/4D scene representation, and Gaussian Splatting. I am interested in how robots can perceive the physical world, build persistent internal models, understand geometry and dynamics, and act intelligently over long horizons. I completed my master's and PhD at Peking University and the Technical University of Munich, where I worked on nonlinear least-squares optimization and Geometric Computer Vision.

My long-term goal is to build next-generation intelligent robots from body to mind, combining robot hardware, spatial understanding, persistent world modeling, and robot action policy learning.

I am actively recruiting motivated PhD and MPhil students, research assistants, and interns to join Embodied Spatial AI Lab at HKUST(GZ), starting from Fall 2026 and Spring 2027.

Join the Lab

About Me

I am currently a Research Fellow at the National University of Singapore and will join HKUST(GZ) as an Assistant Professor in August 2026.

I serve as an Associate Editor for IEEE RA-L, an Area Chair for IROS 2026, and a Lead Organizer of the IROS 2025 Workshop AIR4D.

Scholar

GitHub

Research

Research Program

My research focuses on building embodied spatial intelligence systems that enable robots to perceive, model, reason about, and interact with complex real-world environments.

Embodied Perception

Geometry-aware perception systems that enable robots to localize, map, and understand complex real-world environments through spatially consistent representations.

Spatial World Modeling

Persistent 3D and 4D world models that support long-horizon robot autonomy through geometry-centric tracking, reconstruction, and structured scene representations.

Robot Learning and Action

Learning-based robot systems that connect perception, representation, and reasoning with decision-making and interaction in dynamic environments.

Featured Project

Multi-Sensor Localization, Mapping and Navigation

We are actively recruiting interns and research assistants for a robotics project on multi-sensor fusion, localization, mapping, and autonomous inspection.

Project Overview

We are building a next-generation multi-sensor localization and mapping system for autonomous inspection robots. The goal is to achieve robust, accurate, and real-time localization and high-fidelity 3D reconstruction in large-scale outdoor environments.

The system will integrate multiple sensing modalities, including GNSS/RTK, LiDAR, RGB cameras, IMU, wheel odometry, and Gaussian Splatting-based mapping. The project targets long-term autonomous robot inspection in complex industrial and outdoor environments.

We Are Looking For

We welcome motivated interns and research assistants with strong interests in robotics, SLAM, computer vision, and autonomous systems.

ROS / ROS2 LiDAR RTK / GNSS LIO-SAM FAST-LIO VINS-Fusion ORB-SLAM GTSAM Ceres Kalman Filter Factor Graph 3D Reconstruction

Opportunities

Outstanding students are recommended for 2027 Fall PhD admission at HKUST(GZ). Interested students are encouraged to send me their CV, and a short description of their research interests.

Apply for this Project

News

Recent News

18 June 2026

Three papers are accepted to ECCV 2026 .

21 Feb 2026

One paper is accepted to CVPR 2026 (Findings).

7 Feb 2026

Serving as Associate Editor (AE) for IROS 2026.

31 Jan 2026

One paper is accepted to ICRA 2026 in Alias-free Gaussian Splatting SLAM.

23 Jan 2026

Serving as Session Co-Chair for 3D Computer Vision 1&2 for AAAI-26.

8 Nov 2025

Two oral papers are accepted to AAAI 2026 in RiemanLine Parametrization and Active3D.

11 Oct 2025

Appointed as an Associate Editor of IEEE Robotics and Automation Letters (RA-L).

19 Sep 2025

One paper is accepted to NeurIPS 2025: 4D3R: Motion-Aware Neural Reconstruction and Rendering of Dynamic Scenes from Monocular Videos.

27 Jun 2025

Three papers are accepted to ICCV 2025 in 4D SLAM, Scene Graph Prediction, and Autonomous Driving Gaussian Splatting.

More news

Publications

Selected Publications

*Equal contribution †Corresponding author

AAAI 2026 Oral

RiemanLine: Riemannian Manifold Representation of 3D Lines for Factor Graph Optimization

Y. Li, Z. Yang, K. Tateno, F. Tombari, L. Zhao, G. H. Lee.

Paper Code

AAAI 2026 Oral

Active3D: Active High-Fidelity 3D Reconstruction via Hierarchical Uncertainty Quantification

Y. Li, Y. Li, G. H. Lee.

Project Page

ICCV 2025

4D Gaussian Splatting SLAM

Y. Li, Y. Fang, Z. Zhu, K. Li, Y. Ding, F. Tombari.

Paper Project Page Code

ECCV 2024

GeoGaussian: Geometry-aware Gaussian Splatting for Scene Rendering

Y. Li, C. Lyu, Y. Di, G. Zhai, G. H. Lee, F. Tombari.

Paper Project Page Code

CVPR 2025

Learnable Infinite Taylor Gaussian for Dynamic View Rendering

B. Hu*, Y. Li*, R. Xie, B. Xu, H. Dong, J. Yao, G. H. Lee.

Paper Project Page Code

IEEE RA-L 2024

Open-Structure: a Structural Benchmark Dataset for SLAM Algorithms

Y. Li, Z. Guo, Z. Yang, Y. Sun, L. Zhao, F. Tombari.

Paper Project Page Code

ECCV 2022

E-Graph: Minimal Solution for Rigid Rotation with Extensibility Graphs

Y. Li, F. Tombari

Paper Video

ICRA 2021

ManhattanSLAM: Robust Planar Tracking and Mapping Leveraging Mixture of Manhattan Frames

R. Yunus, Y. Li†, F. Tombari

Paper Code

ICRA 2021

RGB-D SLAM with Structural Regularities

Y. Li, R. Yunus, N. Brasch, N. Navab, F. Tombari.

Paper Code

IEEE RA-L / IROS 2020

Structure-SLAM: Low-drift Monocular SLAM in Indoor Environments

Y. Li, N. Brasch, Y. Wang, N. Navab, F. Tombari.

Paper Code

Remote Sensing Letters 2018

Bundle adjustment method using sparse BFGS solution

Y. Li, S. Fan, Y. Sun, Q. Wang, S. Sun.

Paper

Full Publication List

Openings

Join the EmSAIL Group

I am actively recruiting PhD and MPhil students, research assistants, and interns to join the EmSAIL Group at HKUST(GZ). My research focuses on VLA, Imitation learning, Reinforcement learning, SLAM, 3D/4D scene representation, and Scene understanding.

PhD and MPhil positions in embodied AI, robotics, SLAM, and 3D/4D vision

RA positions in perception, reconstruction, and robot learning

Internships in robot hardware, control, Robot OS, and intelligent systems

Students are encouraged to contact me with their CV and research interests.

What I value

Self-motivation, curiosity, strong execution, and the ability to connect solid theoretical understanding with real-world robotic systems.

Yan Li (李言)

About Me

Research Program

Embodied Perception

Spatial World Modeling

Robot Learning and Action

Multi-Sensor Localization, Mapping and Navigation

Project Overview

We Are Looking For

Recent News

Selected Publications

RiemanLine: Riemannian Manifold Representation of 3D Lines for Factor Graph Optimization

Active3D: Active High-Fidelity 3D Reconstruction via Hierarchical Uncertainty Quantification

4D Gaussian Splatting SLAM

GeoGaussian: Geometry-aware Gaussian Splatting for Scene Rendering

Learnable Infinite Taylor Gaussian for Dynamic View Rendering

Open-Structure: a Structural Benchmark Dataset for SLAM Algorithms

E-Graph: Minimal Solution for Rigid Rotation with Extensibility Graphs

ManhattanSLAM: Robust Planar Tracking and Mapping Leveraging Mixture of Manhattan Frames

RGB-D SLAM with Structural Regularities

Structure-SLAM: Low-drift Monocular SLAM in Indoor Environments

Bundle adjustment method using sparse BFGS solution

Join the EmSAIL Group

Contact