We are glad to share with you that our recent international conference paper “Parameter-Efficient Multi-Source Adaptation for Multimodal Fashion Retrieval” from our MSAAI student Yuanliu Liu, which has been accepted in 2026 International Conference on Artificial Intelligence and Electromechanical Automation (2026 AIEA):
- Yuanliu Liu, Harris Sik-Ho Tsang, Richard Tai-Chiu Hsung, Tony Yulin Zhu, Xiaoxing Yang, Wai-Lun Lo, Ziyin Huang, Yui-Lam Chan, “Parameter-Efficient Multi-Source Adaptation for Multimodal Fashion Retrieval,” International Conference on Artificial Intelligence and Electromechanical Automation (AIEA), Shenzhen, China, Jun. 2026.
Paper Abstract
Online fashion shopping often fails when you try to search for clothes using text descriptions or casual street photos, because standard AI gets confused by the massive difference between messy real-world snaps and polished studio catalog images. To fix this, we created FashLoRA-SigLIP, a smart two-stage AI system. First, we train a robust “fashion brain” on a massive mix of different style datasets so it learns to recognize clothing from any angle. Second, instead of retraining the whole massive model for a new online store, we use a lightweight strategy called Low-Rank Adaptation (LoRA) which updates just 5.6% of the model. When tested on industry-standard datasets, this approach dramatically outperforms standard models, improving the accuracy of finding exact shop items using casual consumer photos, while massively cutting down on computing costs for e-commerce platforms.
In this conference meeting, we obtain the best oral presentation award. Hong Kong Chu Hai College is committed to supporting student research and encouraging them to publish papers in international conferences and journals. More papers coming soon!
The research team members
Hong Kong Chu Hai College
Mr Yuanliu Liu, Master Student in the Department of Computer Science
Dr Harris Sik-Ho Tsang, Assistant Professor in the Department of Computer Science
Dr Richard Tai-Chiu Hsung, Associate Professor in the Department of Computer Science
Dr Tony Yulin Zhu, Assistant Professor in the Department of Computer Science
Dr Xiaoxing Yang, Assistant Professor in the Department of Computer Science
Prof. Wai Lun Lo, Professor and Head of the Department of Computer Science
Shenzhen Polytechnic University
Dr Ziyin Huang, Assistant Professor in Undergraduate School of Artificial Intelligence
The Hong Kong Polytechnic University
Prof. Yui-Lam Chan, Associate Professor and Associate Head in the Department of Electrical and Electronic Engineering
Some photos of the authors and papers
Our proposed FashLoRA-SigLIP, (a) Text-to-image (T2I) retrieval, and (b) view-to-view (V2V) retrieval.
Our proposed two-stage adaptation framework, FashLoRA-SigLIP for multimodal fashion retrieval.
Oral Presentation
Best Oral Presentation Award
