AI Paper Atlas

Research map / Label-efficient and robust learning / Domain adaptation and generalization

Learning to Transfer Learn: Reinforcement Learning-Based Selection for Adaptive Transfer Learning

Linchao Zhu, Sercan . Arık, Yi Yang, Tomas Pfister

ECCV 2020 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Statistical alignment and hypothesis testing · 12 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
add L2TL RL-based source class weightingfine-tuning · Birdsnap (ImageNet source)top-1 accuracy: 77.1 -> 78.1Improved
add L2TL RL-based source class weightingfine-tuning · Stanford Cars (ImageNet source)top-1 accuracy: 92.0 -> 92.6Improved
add L2TL RL-based source class weightingfine-tuning · FGVC Aircraft (ImageNet source)top-1 accuracy: 88.2 -> 89.1Improved
use L2TL instead of DATL class-prior matchingDATL · Birdsnap (JFT-300M source)top-1 accuracy: 81.7 -> 82.4Improved
add L2TL with large JFT-300M sourcefine-tuning · Birdsnap (JFT-300M source)top-1 accuracy: 74.9 -> 82.4Improved
add L2TL for MNIST to SVHN transferfine-tuning · SVHN (MNIST source, 600 train images)top-1 accuracy: 71.7 -> 75.2Improved
add L2TL in low-shot regime (5 samples per class)fine-tuning · Stanford Cars (5 samples per class)top-1 accuracy: 40.0 -> 46.5Improved
add L2TL with 20 samples per classfine-tuning · Stanford Cars (20 samples per class)top-1 accuracy: 84.9 -> 86.1Improved
add L2TL on dissimilar-domain texture targetfine-tuning · DTD split 1 (full training set)top-1 accuracy: 70.3 -> 72.0Improved
add L2TL on DTD with 10 examples per classfine-tuning · DTD split 1 (10 examples per class)top-1 accuracy: 55.0 -> 60.1Improved
add L2TL on chest X-ray targetfine-tuning · CheXpert (Mean)AUC: 89.3 -> 90.0Improved
use ImageNet fine-tuning instead of random initializationrandom initialization · DTD split 1 (full training set)top-1 accuracy: 57.4 -> 70.3Improved

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