Yuetian Chen

Graduate Student

Joined department: Fall 2024

My research area is natural language processing, with a focus on developing and deploying controllable and interactive content generation based on language models. This involves understanding open-world, complex, and multi-modal data, and establishing a further application for deep learning in computational creativity.

Recently, I have been dedicated to researching security issues of foundational models, including: (1) examining various Membership Inference Attack methods in computer vision tasks, such as Likelihood Ratio Attack and Threshold Attack, dynamically and from the perspective of individual data samples; (2) Adversarial Attacks against Large Language Models (LLMs), like Prompt-level Jail Break and Data Poisoning.

Additionally, I utilize artificial intelligence to deploy human-centric multimodal agent robots, enhancing their interactivity and multi-functionality in diverse environments.