Papers by Zhang Liujie
HiSVD: Principled Low-Rank Approximation of LLMs via Hierarchical Modeling of Information Capacity and Spectral Structure (2026.acl-long)
Copied to clipboard
| Challenge: | Existing methods for generating layer importance ignore the fine-grained influence of spectral distribution shape. |
| Approach: | They propose a hierarchical rank allocation framework with two stages to address this gap . they propose SVD-based lowrank approximation that exploits spectral heterogeneity . |
| Outcome: | Experiments show that HiSVD outperforms state-of-the-art methods on LLMs . |
MMAG: Multimodal Learning for Mucus Anomaly Grading in Nasal Endoscopy via Semantic Attribute Prompting (2025.emnlp-main)
Copied to clipboard
| Challenge: | Accurate grading of rhinitis severity relies heavily on the characterization of key secretions, notably clear nasal discharge (CND) and purulent nasal secretion (PUS). |
| Approach: | They propose a framework that integrates structured prompts with rank-aware vision-language modeling for joint detection and grading. |
| Outcome: | The proposed model improves AUC and F1 scores on CND and PUS datasets by 6.31% and 4.79%. |