AI Paper Atlas

Research map / Probabilistic and scientific modeling / Molecules and proteins

Determining the Number of Latent Factors in Statistical Multi-Relational Learning

Chengchun Shi, Wenbin Lu, Rui Song

JMLR 2019 · Probabilistic and scientific modeling › Molecules and proteins · approach: Sparse graphical and tensor models · 8 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
use proposed IC instead of BIC-type criterionlogistic RESCAL · Simulation Settings I-VItrue-model selection rate: BIC 0.00 vs IC0 0.97 (n=100, K=3, s0=2)Improved
add finite-sample correction tau_alpha with alpha=0.5 (IC0 -> IC0.5) (ablation)IC0 · Simulation n=50, K=50, s0=8TP: 0.17 -> 0.76Improved
use alpha=1 correction instead of alpha=0 for large K (ablation)IC0 · Simulation n=50, K=20, s0=8TP: 0.73 -> 0.79Improved
increase alpha from 0 to 1 at small K (ablation)IC · Simulation n=100, K=3TP: 0.97 -> 0.97 (s0=2)No clear effect
increase alpha from 0 to 1 at K=10, s0=8 (ablation)IC0 · Simulation n=50, K=10, s0=8TP: 0.69 -> 0.60Hurt
use IC instead of BIC for factor number selectionlogistic RESCAL · Social Evolutionnumber of factors selected: 9 vs 12 factorsImproved
increase number of factors s from 1 to 10logistic RESCAL · Social EvolutionAUC: 0.7201 -> 0.9518Improved
increase number of factors s from 10 to 12logistic RESCAL · Social EvolutionAUC: 0.9518 -> 0.9467Hurt

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