Research map / Architectures and efficiency / RNNs and normalization
SyMerge: From Non-Interference to Synergistic Merging via Single-Layer Adaptation
Aecheon Jung, Seunghwan Lee, Dongyoon Han, Sungeun Hong
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| jointly train merging coefficients with the task-specific layer (vs layer only) (ablation) | SyMerge · 20 vision tasks, ViT-B/32 | average accuracy: 75.0 -> 88.6 | Improved |
| jointly train task-specific layer with coefficients (vs coefficients only) (ablation) | SyMerge · 8 vision tasks, ViT-B/32 | average accuracy: 84.6 -> 90.1 | Improved |
| adapt multiple layers at once instead of a single layer (ablation) | SyMerge · 8 vision tasks, ViT-B/32 | accuracy | Hurt |
| train a single internal transformer layer instead of the classifier (ablation) | SyMerge · 8 vision tasks, ViT-B/32 | accuracy | No clear effect |
| use SyMerge-learned classifiers with a Task Arithmetic merged encoder | Task Arithmetic · 8 vision tasks, ViT-B/32 | merged accuracy: 69.1 -> 79.6 (+10.5) | Improved |
| use SyMerge-learned classifiers with an AdaMerging encoder | AdaMerging · 8 vision tasks, ViT-B/32 | merged accuracy: 80.1 -> 87.7 (+7.6) | Improved |
| keep SyMerge merged encoder but swap in zero-shot classifier (ablation) | SyMerge · 8 vision tasks, ViT-B/32 | average accuracy: 69.1 -> 82.4 | Improved |
| self-labeling with expert predictions instead of entropy minimization (ablation) | SyMerge · 8 vision tasks, ViT-B/32 | Spearman correlation with ground-truth loss | Improved |
| sequential per-task update strategy (ablation) | AdaMerging · 8 vision tasks, ViT-B/32 | average accuracy: 80.1 -> 81.6 | Improved |
| use SyMerge for merging models from different initializations | SyMerge · EuroSAT+DTD disjoint basins | average accuracy: 6.9 -> 79.1 | Improved |
| supervise with a pretrained/individual interpolation lambda 0.8 instead of individual models (ablation) | SyMerge · 8 vision tasks, ViT-B/32 | average accuracy: 90.1 -> about 90.2 | Improved |
| jointly train coefficients with classifier (vs classifier only) for init robustness (ablation) | SyMerge · 8 vision tasks, ViT-B/32 | accuracy vs initial coefficient | Improved |
Closest papers by meaning
- Modeling Multi-Task Model Merging as Adaptive Projective Gradient Descent (ICML 2025 · similarity 0.90)
- Twin-Merging: Dynamic Integration of Modular Expertise in Model Merging (NeurIPS 2024 · similarity 0.90)
- AdaMerging: Adaptive Model Merging for Multi-Task Learning (ICLR 2024 · similarity 0.90)
- Whoever Started the interference Should End It: Guiding Data-Free Model Merging via Task Vectors (ICML 2025 · similarity 0.88)
- Towards Minimizing Feature Drift in Model Merging: Layer-wise Task Vector Fusion for Adaptive Knowledge Integration (NeurIPS 2025 · similarity 0.88)
- EMR-Merging: Tuning-Free High-Performance Model Merging (NeurIPS 2024 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
