AI Paper Atlas

Research map / Label-efficient and robust learning / OOD and anomaly detection

PIDForest: Anomaly Detection via Partial Identification

Parikshit Gopalan, Vatsal Sharan, Udi Wieder

NeurIPS 2019 · Label-efficient and robust learning › OOD and anomaly detection · approach: Anomaly detection with contaminated data · 4 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 PIDForest instead of iForestAnomaly detection · ThyroidAUC: 0.819 ± 0.013 -> 0.876 ± 0.013Improved
Use PIDForest instead of PCAAnomaly detection · MammographyAUC: 0.886 -> 0.840 ± 0.010Hurt
Use PIDForest instead of iForestAnomaly detection · smtpAUC: 0.908 ± 0.003 -> 0.923 ± 0.003Improved
Use PIDForest instead of kNNAnomaly detection · NYC taxicabAUC: 0.697 -> 0.564 ± 0.004Hurt
Add 50 random dimensionsPIDForest · MammographyAUC: 2% dropHurt
Add 50 random dimensionsiForest · MammographyAUC: 15% dropHurt
Add 50 random dimensionsPIDForest · ThyroidAUC: 13% dropHurt
Add 50 random dimensionsiForest · ThyroidAUC: 20% dropHurt
Use leaf depth instead of leaf sparsity (ablation)PIDForest · ThyroidAUC: 0.876 -> 0.847Hurt
Use leaf depth instead of leaf sparsity (ablation)PIDForest · MammographyAUC: 0.840 -> 0.783Hurt
Set the number of buckets to 2 instead of 3, 4, 5, or 6 (ablation)PIDForest · Synthetic time seriesprecision-recallHurt
Increase noisy-dimension scale from [-2,2] to [-10,10]PIDForest · Gaussian mixture with random noisetrue anomalies in top 100No clear effect

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