AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Learning Neural Bag-of-Matrix-Summarization with Riemannian Network

Hong Liu, Jie Li, Yongjian Wu, Rongrong Ji

AAAI 2019 · Label-efficient and robust learning › Recognition and long-tail · approach: Correspondence and causal representation learning · 7 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 data summarization with metric learning (BoMS-2) to SPDNetSPDNet · AFEWaccuracy: 34.23% -> 38.81%Improved
add BoMS-2 to SPDNetSPDNet · HDM05accuracy: 61.45% -> 71.79%Improved
add BoMS-2 to SPDNetSPDNet · BCI IV 2aaccuracy: 55.59% -> 65.20%Improved
add BoMS-2 to SPDNetSPDNet · YTCaccuracy: 89.01% -> 89.81%Improved
remove metric learning (BoMS-ML) (ablation)BoMS-2 · AFEWaccuracy: 34.23% -> 38.81%Improved
remove metric learning (BoMS-ML) (ablation)BoMS-2 · YTCaccuracy: 83.04% -> 89.81%Improved
remove divergence loss (BoMS-D, lambda2=0) (ablation)BoMS-2 · AFEWaccuracy: 36.93% -> 38.81%Improved
remove divergence loss (BoMS-D) (ablation)BoMS-2 · BCI IV 2aaccuracy: 65.08% -> 65.20%No clear effect
remove data summarization (BoMS-A) (ablation)BoMS-2 · YTCaccuracy: 80.02% -> 89.81%Improved
remove data summarization (BoMS-A) (ablation)BoMS-2 · HDM05accuracy: 68.61% -> 71.79%Improved
use metric-learned Log-Euclidean measure (BoMS-2) instead of plain LEM (BoMS-1) (ablation)BoMS · AFEWaccuracy: 35.04% -> 38.81%Improved
change summarized set size (ablation)BoMS-2 · AFEWaccuracy: 35.3 -> 38.8 -> 34.5Mixed

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