AI Paper Atlas

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

Zero-Shot Deep Domain Adaptation

Kuan-Chuan Peng, Ziyan Wu, Jan Ernst

ECCV 2018 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Geometric alignment and synthetic-to-real adaptation · 95 citations (OpenAlex)

Read the paper

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 ZDDA2 training with task-irrelevant dual-domain pairsLeNet source-only classifier · MNIST->MNIST-Maccuracy (%): 39.04 -> 94.84Improved
add ZDDA2 with task-irrelevant pairsLeNet source-only classifier · NIST->NIST-Maccuracy (%): 8.59 -> 34.25Improved
use a more semantically similar task-irrelevant dataset (EMNIST rather than Fashion-MNIST)ZDDA2 · MNIST->MNIST-Maccuracy (%): 73.15 -> 94.84Improved
use MNIST (S 0.049) rather than EMNIST (0.178) as T-I dataZDDA2 · Fashion-MNIST->Fashion-MNIST-Maccuracy (%): 51.55 -> 65.30Mixed
add ZDDA2 training with T-I RGB-D pairs, testing on RGBGoogleNet · SUN RGB-D 10 scenesaccuracy (%): 31.58 -> 43.16Improved
add ZDDA2 training with T-I RGB-D pairs, testing on depthGoogleNet · SUN RGB-D 10 scenesaccuracy (%): 51.84 -> 59.74Improved
add step 3 (joint network without L2 constraint, ZDDA3) to ZDDA2 (ablation)ZDDA2 · SUN RGB-D 10 scenes (depth test)accuracy (%): 59.74 -> 63.16Improved
add step 3 (ZDDA3) to ZDDA2 (ablation)ZDDA2 · SUN RGB-D 10 scenes, AlexNetaccuracy (%): 51.05 -> 51.05No clear effect
add step 3 (ZDDA3) to ZDDA2 (ablation)ZDDA2 · SUN RGB-D 10 scenes, 5-fold splitsmean accuracy (%): 56.89 -> 58.37Improved
add ZDDA2 on a different base architecture (SqueezeNet), depth testSqueezeNet_v1.1 · SUN RGB-D 10 scenesaccuracy (%): 48.68 -> 56.32Improved
add ZDDA2 on AlexNet, depth testAlexNet · SUN RGB-D 10 scenesaccuracy (%): 49.74 -> 51.05Improved

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