Brief Biography
Introduction
I am a Ph.D. candidate in Software Engineering at Dalian University of Technology (DLUT), expecting to graduate in June 2025. I am affiliated with the Institute of Geometric Computing and Intelligent Media Technology and the Key Laboratory for Ubiquitous Network and Service Software of Liaoning Province. My research is supervised by Prof. Risheng Liu (National Excellent Young Scholar).
My research centers on the development of reliable and efficient intelligent optimization algorithms, bridging the domains of machine learning and multimedia. I focus on advancing hierarchical learning tasks, such as Hyperparameter Optimization (HO), Meta-Learning, Neural Architecture Search (NAS), and Generative Adversarial Networks (GANs), alongside coupled vision applications like Semantic Segmentation (SS), Fusion, and SuperResolution (SR). I also design and implement tailored optimization methods to address key Trustworthy Machine Learning problems, including Adversarial Training and Black-box Transfer Attacks. Currently, I am deeply investigating the intricate learning challenges posed by large-scale machine learning models, such as fine-tuning, unlearning, and pressing security concerns, including multi-modal attacks and LLM Jailbreaking.
I have authored 10 SCI/EI papers and have 6 additional papers under review. My first-author publications include contributions to NeurIPS, ICML, IJCAI, and ICME, among others. I have also contributed to two international patents. Additionally, I serve as a reviewer for leading conferences and journals, including CVPR, ACM MM, AAAI, and IJCAI.
Beyond research, I actively participate in more than 4 national key R&D and NSFC-funded projects, designing and implementing optimization-driven solutions for image processing and machine learning tasks.
Postdoctoral Opportunity
I am currently seeking postdoctoral opportunities in areas related to bi-Level optimization, trustworthy machine learning and multimedia. I am particularly interested in interdisciplinary collaborations that explore novel optimization methods and interesting real-world applications like LLM finetuning and security issues. If you are interested in these or related areas, I would be delighted to discuss potential collaborations. Please feel free to contact me at liuyaohua.918@gmail.com.
Recent News
- [11/2024] One paper titled “Enhancing Images with Coupled Low-Resolution and Ultra-Dark Degradations: A Tri-level Learning Framework” got accepted at ACM MM 2024.
- [10/2024] One paper titled “A Dual-Stream-Modulated Learning Framework for Illuminating and Super-Resolving Ultra-Dark Images” got accepted at IEEE TNNLS 2024.
- [09/2024] One paper titled “Collaborative Brightening and Amplification of Low-Light Imagery Via Bi-Level Adversarial Learning” got accepted at Pattern Recognition (PR) 2024.
- [08/2024] One paper titled “Advancing Generalized Transfer Attack with Initialization Derived Bilevel Optimization and Dynamic Sequence Truncation” got accepted at IJCAI 2024.
- [05/2023] One paper titled “Averaged method of multipliers for bi-level optimization without lower-level strong convexity” got accepted at ICML 2023.
- [8/2023] One paper titled “PEARL: Preprocessing Enhanced Adversarial Robust Learning of Image Deraining for Semantic Segmentation” got accepted at ACM MM 2023.
For more info
For more details, please refer to my GitHub or contact me here (mail: liuyaohua.918@gmail.com, Phone: +87 18742014196, WeChat: me_for_u, QQ: 919193083).