Showing cs.LGShow all
3 papers · 1 filter
cs.LG2025
Towards Minimizing Feature Drift in Model Merging: Layer-wise Task Vector Fusion for Adaptive Knowledge Integration
Wenju Sun, Qingyong Li, Wen Wang +3
Multi-task model merging aims to consolidate knowledge from multiple fine-tuned task-specific experts into a unified model while minimizing performance degradation. Existing method…
cs.LG2025
Task Arithmetic in Trust Region: A Training-Free Model Merging Approach to Navigate Knowledge Conflicts
Wenju Sun, Qingyong Li, Wen Wang +2
Multi-task model merging offers an efficient solution for integrating knowledge from multiple fine-tuned models, mitigating the significant computational and storage demands associ…
cs.LG2022
Exemplar-free Class Incremental Learning via Discriminative and Comparable One-class Classifiers
Wenju Sun, Qingyong Li, Jing Zhang +3
The exemplar-free class incremental learning requires classification models to learn new class knowledge incrementally without retaining any old samples. Recently, the framework ba…