AI Paper Atlas

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

Model Fusion via Retrofitting

Phoomraphee Luenam, Andreas Spanopoulos, Amit Sant, Sotiris Anagnostidis, Thomas Hofmann, Sidak Pal Singh

ICML 2026 · Label-efficient and robust learning › Zero-/few-shot and meta-learning · approach: Zero-shot robustness and transfer · 0 citations (OpenAlex)

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
use KF-Gradient instead of Transformer OTFusionViT · CIFAR-100 sharded 4-way, zero-shottest accuracy: 1.2 -> 41.4Improved
use HF-Gradient instead of KDViT · CIFAR-100 sharded 2-way, zero-shottest accuracy: 41.0 -> 54.5Improved
use KF-Gradient instead of KDViT · CIFAR-100 sharded 4-waytest accuracy: 47.0 -> 41.4Hurt
use KF-Linear instead of ensembleVGG · BloodMNIST sharded 2-waytest accuracy: 88.4 -> 91.0Improved
use KF-Gradient (25k samples) instead of ZipIt! (25k)ResNet-50 · ImageNet-1k two subsets, jointtest accuracy: 37.7 -> 58.9Improved
increase fusion samples for KF-Gradient from 1k to 25k (ablation)ResNet-50 · ImageNet-1k jointtest accuracy: 25.1 -> 58.9Improved
add data augmentation to 1k fusion samples (ablation)KF-Gradient VGG11 · CIFAR-10 non-IID, zero-shottest accuracy: 73.2 -> 74.5 (uniform)Improved
use Conductance or DeepLIFT importance instead of uniform with 1k samples (ablation)KF-Gradient VGG11 · CIFAR-10 non-IID, zero-shottest accuracy: 73.2 -> 71.7 / 71.4Hurt
use Conductance or DeepLIFT importance instead of uniform with 1k samples plus augmentation (ablation)KF-Gradient VGG11 · CIFAR-10 non-IID, zero-shottest accuracy: 74.5 -> 75.3 / 75.2Improved
use KF-Gradient instead of Transformer OTFusionViT · CIFAR-100 full data 2-way, zero-shottest accuracy: 4.3 -> 63.0Improved
use KF-Gradient instead of Transformer OTFusion after fine-tuningViT · CIFAR-100 full data 2-waytest accuracy: 74.0 -> 75.4Improved
use KF-Gradient instead of Transformer OTFusion after fine-tuningViT · Tiny-ImageNet 2-waytest accuracy: 53.8 -> 54.2Improved

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