AI Paper Atlas

Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning

Learn to Merge: Meta-Learning for Adaptive Multi-Task Model Merging

Jun Chen, Qin Zhang, Weizhi Zhang, Xiao Luo, Philip S. Yu, Ziyue Qiao

ICML 2026 · Label-efficient and robust learning › Zero-/few-shot and meta-learning · approach: Meta-augmentation and meta-objectives

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 changedWhereMeasuredEffect
meta-learn merging coefficients with MetaMerging versus AdaMergingViT-B/32 merging · 8 vision tasksavg accuracy: 80.1 -> 87.2Improved
MetaMerging versus Surgery on ViT-L/14ViT-L/14 merging · 8 vision tasksavg accuracy: 92.3 -> 93.4Improved
MetaMerging versus Weight Averaging on GPT-2GPT-2 merging · GLUE 7 tasksavg accuracy: 56.1 -> 76.9Improved
add task-specific adapters to the meta-learned unified model (ablation)MetaMerging · ViT-B/32 8 tasksavg accuracy: 67.8 -> 87.2Improved
add the same adapters to Weight Averaging merged model (ablation)Weight Averaging · ViT-B/32 8 tasksavg accuracy: 65.8 -> 80.0Improved
meta-learn coefficients instead of AdaMerging coefficients, both with adapters (ablation)unified model + adapters · ViT-B/32 8 tasksavg accuracy: 86.1 -> 87.2Improved
meta-learned unified model before adapters is weaker than AdaMerging before adapters (ablation)MetaMerging · ViT-B/32 8 tasks, no adaptersavg accuracy: 80.1 -> 67.8Hurt
use one inner adapter step instead of zero in meta-learning (GPT-2) (ablation)MetaMerging · GPT-2 GLUEavg accuracy: 74.9 -> 76.4Improved
increase inner steps from 1 to 5 or 10 (ablation)MetaMerging · GPT-2 GLUEavg accuracy: 76.4 -> 75.7 (5 steps), 75.8 (10 steps)Hurt
set inner step size to 1 instead of 0.1 (ablation)MetaMerging · ViT-B/32 8 tasksavg accuracy: 87.15 -> 86.74Hurt
set meta step size to 1 instead of 0.01 (ablation)MetaMerging · ViT-B/32 8 tasksavg accuracy: 87.15 -> 85.74Hurt
use MetaMerging instead of multi-task learning for training costViT-B/32 merging · 8 vision tasksrunning time: 15h 53m -> 1h 23mImproved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy