Qi Fang(方淇)

I am a first-year CS Master student at University of Toronto.

My research interests are in Computer Vision, AI Security, and Robotics. Specifically, to make self-driving cars robust under the conditions of adverse weather and adversarial attacks.

I am also interested in LLM x Robotics, Embodied AI, HCI, and AI for Social Good.

I was fortunately in research collaboration at AI4CE LAB, New York University, supervised by Prof. Chen Feng and Dr. Felix Juefei Xu. Previously I interned at iDLAB, Tsinghua University, advised by Prof. Wenjun Wang, where I focused on human behavior interacting with self-driving vehicles.

Feel free to contact me by email: qifang at cs.toronto.edu

Gmail  |  Google Scholar  |  Github

We proposed a novel probability-guaranteed sampling method inspired by RANSAC to achieve a trust-worthy group of benign agents, outperforming adversarial training Method by 73%.

We proposed a new LiDAR perception task, Occupancy Completion and Forecasting (OCF), combining scene completion and forecasting in autonomous driving, and evaluates it using the newly curated OCFBench dataset.

We validated a key indicator of relative position between the ego vehicle and surrounding vehicles having important effects on take-over performance and driving safety.

"To augment the performance of lidar perception under challenging meteorological conditions. Focusing on noise-removal and noise-utilization methods."

"To explore the potential of reverse adversarial attacks to realize robust vision as well as robust 'objects' to current deep vision models."

Research Topics:
[1] Multi-agent Collaboration / [2] Adversarial Machine Learning / [3] Robust Robot Vision

[1][2] -> Among Us, accepted to ICCV 23'
[2] -> Reverse Adversarial Learning
[3] -> 3D Perception in Adverse Weather Conditions

Tasks: Design, develop, and test autonomous driving strategies for DiDi’s level-4 robotaxis to optimize the design of algorithms, data science models, and autonomous features in terms of their reliability, safety, and efficiency

Tasks: Research, quality assurance, user testing, evaluation and optimization of XPeng’s Valet Parking Assist (VPA) feature

Research Topics: Human factors and ergonomics in conditional automated driving. -> AHFE 2021 Paper

Major Courses: Operating System / Database System / Discrete Mathematics / Machine Learning / Data Structures / Compilers / Computer Networks