Yunhao Luo

I am a visiting student researcher at Georgia Tech advised by Prof. Danfei Xu. I also work closely with Prof. Yilun Du. I received my M.S. from Brown CS in 2024, where I worked with Prof. Chen Sun. Before that, I worked with Prof. Lin Wang at HKUST(GZ) and Prof. Jianguo Zhang at SUSTech on computer vision. I received my B.E. in CS from SUSTech in 2022.

My research goal is to develop intelligent agents that can freely interact with the physical world and bring a positive impact on people's lives. To this end, my interests lie at the intersection of computer vision, decision-making, and robotics. I seek to build intelligence agents in a more efficient and scientific way by drawing inspiration from human cognitive biases and make them more reliable, lightweight, cost-effective, and accessible to users. Currently, I am working on leveraging generative AI to enable robust and generalizable decision-making.

Outside academia, I spend my time playing sports and music, and hanging out with friends.

Publications ( show selected / show by date / show by topic )

Topics: Decision Making / Computer Vision (* indicates equal contribution and † indicates equal advising)

Grounding Video Models to Actions through Goal Conditioned Exploration
Yunhao Luo, Yilun Du

ArXiv Preprint / Website / Paper / Code

Potential Based Diffusion Motion Planning
Yunhao Luo, Chen Sun, Joshua B. Tenenbaum, Yilun Du

ICML 2024 / Website / Paper / Code / Colab

A Good Student is Cooperative and Reliable: CNN-Transformer Collaborative Learning for Semantic Segmentation
Jinjing Zhu, Yunhao Luo, Xu Zheng, Hao Wang, Lin Wang

ICCV 2023 / Paper

Look at the Neighbor: Distortion-aware Unsupervised Domain Adaptation for Panoramic Semantic Segmentation
Xu Zheng, Tianbo Pan, Yunhao Luo, Lin Wang

ICCV 2023 / Paper / Code

Transformer-CNN Cohort: Semi-supervised Semantic Segmentation by the Best of Both Students
Xu Zheng, Yunhao Luo, Chong Fu, Kangcheng Liu, Lin Wang

ICRA 2024 / Paper

Distilling Efficient Vision Transformers from CNNs for Semantic Segmentation
Xu Zheng, Yunhao Luo, Pengyuan Zhou, Lin Wang

Pattern Recognition / Paper

Experience
Visiting Student Researcher, Georgia Tech
Sep 2024 -- Now

Working on robot learning and generative model, advised by Prof. Danfei Xu, School of Interactive Computing.

Software Engineer Intern, Tencent
Jul 2021 -- Sep 2021

Worked for WeCom, a work chat and office tool with 180M+ active users, mobile app development team. Implemented 10+ features for user access control logics. User interface design, code refactoring.

Exchange Student, UC Berkeley
Jan 2020 -- May 2020

CS 170: Efficient Algorithms and Intractable Problems
CS 61C: Great Ideas in Computer Architecture (Machine Structures)
Math 55: Discrete Mathematics

Summer Workshop, National University of Singapore
Jul 2019 -- Aug 2019

Teaching
Teaching Assistant, SUSTech
Sep 2020 -- Jul 2022

  • CS203B: Data Structures and Algorithm Analysis B,   Spring 2022
  • CS102A: Introduction to Computer Programming,   Fall 2021
  • CS307: Principles of Database Systems,   Spring 2021
  • CS205: C/C++ Programming,   Fall 2020

    • Projects
    Benchmarking the Vulnerability of Skin Cancer Classification
    Jul 2020 -- Apr 2021

    Co-developed a codebase to benchmark the adversarial vulnerability of deep learning based skin cancer classifiers. Evaluated on dermoscopic lesion images with attack methods including FGSM, PGD, and DeepFool.

    SQL Code Evaluation Web Platform
    Sep 2020 -- Jan 2021

    Led the development of a web platform for evaluating SQL code. Based on Java, Spring Boot, Spring Security, PostgreSQL, Redis, etc. Project successfully deployed online for grading student assignments for SUSTech CS307 Principles of Database Systems (∼200 students enrollment).