Zijiao Yang

Zijiao Yang (杨子蛟 in Chinese) is a Ph.D. student at Oregon State University, advised by Prof. Stefan Lee. He earned Master’s degree in computer science at University of Colorado at Boulder.

His research interest is ultimately driven by the mystery of human intelligence, and he believes achieving a “deep” level of language grounding is one of the key steps. He has worked on semantic role labeling during Master’s. Currently, he takes vision-language navigation tasks as a platform to study groundness of natural language onto vision and actions.

News

Feb 27, 2023	Our work: “Behavioral Analysis of Vision-and-Language Navigation Agents” has been accepted to CVPR 2023 (25%)
May 11, 2021	Created personal website!

Selected publications

WACV

Hijacking Vision-and-Language Navigation Agents with Adversarial Environmental Attacks

Yang, Zijiao, Shi, Xiangxi, Slyman, Eric, and Lee, Stefan

IEEE/CVF Winter Conference on Applications of Computer Vision 2025

Abs PDF

Assistive embodied agents that can be instructed in natural language to perform tasks in open-world environments have the potential to significantly impact labor tasks like manufacturing or in-home care – benefiting the lives of those who come to depend on them. In this work, we consider how this benefit might be hijacked by local modifications in the appearance of the agent’s operating environment. Specifically, we take the popular Vision-and-Language Navigation (VLN) task as a representative setting and develop a whitebox adversarial attack that optimizes a 3D attack object’s appearance to induce desired behaviors in pretrained VLN agents that observe it in the environment. We demonstrate that the proposed attack can cause VLN agents to ignore their instructions and execute alternative actions after encountering the attack object – even for instructions and agent paths not considered when optimizing the attack. For these novel settings, we find our attacks can induce early-termination behaviors or divert an agent along an attacker-defined multi-step trajectory. Under both conditions, environmental attacks significantly reduce agent capabilities to successfully follow user instructions.
CVPR

Behavioral Analysis of Vision-and-Language Navigation Agents

Yang, Zijiao, Majumdar, Arjun, and Lee, Stefan

Computer Vision and Pattern Recognition 2023

Abs HTML PDF

To be successful, Vision-and-Language Navigation (VLN) agents must be able to ground instructions to actions based on their surroundings. In this work, we develop a methodology to study agent behavior on a skill-specific basis – examining how well existing agents ground instructions about stopping, turning, and moving towards specified objects or rooms. Our approach is based on generating skill-specific interventions and measuring changes in agent predictions. We present a detailed case study analyzing the behavior of a recent agent and then compare multiple agents in terms of skill-specific competency scores. This analysis suggests that biases from training have lasting effects on agent behavior and existing models have grounding simple referring expressions. Our comparisons between models shows that skill-specific scores correlate with improvements in overall VLN task performance.