AI Paper Atlas

Research map / Label-efficient and robust learning / Semi-supervised learning

Distribution-Based Semi-Supervised Learning for Activity Recognition

Hangwei Qian, Sinno Jialin Pan, Chunyan Miao

AAAI 2019 · Label-efficient and robust learning › Semi-supervised learning · approach: Safe and realistic semi-supervised learning · 14 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 unlabeled-data manifold warping of the RKHS (DSSL over SMMAR) (ablation)SMMAR · Skodamicro-F1: 93.2 -> 98.8Improved
add unlabeled-data manifold warping of the RKHS (DSSL over SMMAR) (ablation)SMMAR · HCImicro-F1: 82.2 -> 99.9Improved
add unlabeled-data manifold warping of the RKHS (DSSL over SMMAR) (ablation)SMMAR · WISDMmicro-F1: 20.5 -> 56.5Improved
distribution embedding instead of handcrafted vector features with semi-supervised SVMLapSVM · WISDMmicro-F1: 40.1 -> 56.5Improved
increase ECDF descriptors d from 5 to 45ECDF+SVM · Skodamicro-F1: 84.2 -> 65.7Hurt
increase ECDF descriptors d from 5 to 45ECDF+SVM · HCImicro-F1: 67.7 -> 68.6No clear effect
increase weight r on unlabeled data beyond 1 (ablation)DSSL · WISDMmicro-F1Hurt
add more unlabeled training datasemi-supervised methods · WISDMmicro-F1Improved
use random Fourier features approximation (R-DSSL) instead of exact kernels (ablation)DSSL · WISDMmicro-F1Hurt
use semi-supervised logistic regression instead of graph-based methodsSSKLR · Skodamicro-F1: 89.7 -> 25.4Hurt

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