AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

Guided Visual Exploration of Relations in Data Sets

Kai Puolamäki, Emilia Oikarinen, Andreas Henelius

JMLR 2021 · Vision-language and multimodal › Human and embodied modeling · approach: Scene graph generation · 5 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 hypothesis-specific projection (matching the hypothesis pair) instead of PCA or ICA projectionProposed view finding · German data, gaingain G(v,H): 8.875 vs PCA 4.211 for H_E,{t}Improved
use PCA instead of the proposed projection with no background knowledgeProposed view finding · German data, H_E,emptygain G(v,H): 8.831 vs 8.831No clear effect
use ICA objective instead of the proposed projectionProposed view finding · German data, gaingain G(v,H): 8.831 -> 0.005Hurt
remove half of the rows (Delta n=200) from the dataProposed framework · Stability, sigma=0mean relative error in gain: 0.000 -> 0.021No clear effect
add Gaussian noise with sigma 5 to 10Proposed framework · Stability, Delta n=0mean relative error in gain: 0.000 -> 0.280, 0.358Hurt
add Gaussian noise with sigma=1Proposed framework · Stability, Delta n=0mean relative error in gain: 0.000 -> 0.049No clear effect
increase data size from n=500 to n=10000 at m=200Proposed framework · Scalabilityview time (s): 1.74 -> 26.37Hurt

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