AI Paper Atlas

Research map / Learning theory and algorithms / Matrix and tensor methods

Robustness Can Be Cheap: A Highly Efficient Approach to Discover Outliers under High Outlier Ratios

Siqi Wang, En Zhu, Xiping Hu, Xinwang Liu, Qiang Liu, Jianping Yin, Fei Wang

AAAI 2019 · Learning theory and algorithms › Matrix and tensor methods · approach: Sparse and robust PCA · 6 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 LEOD-basic/fast low-rank framework instead of other outlier detectors (best other: RKDE)LEOD / UOCL · SVHN under high outlier ratiof1-score: at most 29.5% f1-score gainImproved
Use low-rank similarity (LEOD) instead of full similarity matrix (UOCL)LEOD-basic · Caltech101, outlier ratio 0.6f1-score: 82.6 -> 85.3Improved
Use LEOD-fast instead of UOCLLEOD-fast · SVHN, outlier ratio 0.6time (sec.): 3439.0 -> 3.4Improved
Use LEOD-basic instead of UOCLLEOD-basic · SVHN, outlier ratio 0.6peak memory (GiB): 39.06 -> 0.36Improved
Use LEOD-fast instead of LEOD-basic (closed-form alpha update)LEOD · SVHN, outlier ratio 0.6time (sec.): 247.2 -> 3.4Improved
Use LEOD-basic instead of LEOD-fastLEOD · SVHN, outlier ratio 0.6peak memory (GiB): 0.40 -> 0.36Improved
Naively low-rank approximate (I + gamma1 L) instead of sparsifying L (ablation)LEOD · Caltech101 / Cifar-10f1-score: at most 6% performance lossHurt
Vary lambda from 0.01 to 10.0 (ablation)LEOD-fast · Caltech101f1-scoreNo clear effect
Large lambda = 100 at outlier ratio 60% (ablation)LEOD-fast · Caltech101, outlier ratio 0.6f1-score: at most 4% f1 gainImproved
Large lambda = 100 at outlier ratio <= 40% (ablation)LEOD-fast · Caltech101f1-scoreHurt
Very small lambda = 0.01 at outlier ratio 60% (ablation)LEOD-fast · Caltech101, outlier ratio 0.6f1-scoreHurt
Vary m from n/5 to n/20 and l from n/10 to n/30 (ablation)LEOD · Caltech101f1-score: variance within 3%No 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