AI Paper Atlas

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

DICE: Leveraging Sparsification for Out-of-Distribution Detection

Yiyou Sun, Yixuan Li

ECCV 2022 · Label-efficient and robust learning › OOD and anomaly detection · approach: Provable and feature-based OOD detection · 182 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 DICE weight sparsification to the energy scoreEnergy score · ImageNet-1k ID, avg of 4 OOD setsFPR95: 58.41 -> 34.75Improved
add DICE weight sparsification to the energy scoreEnergy score · CIFAR-100 ID, avg of 6 OOD setsFPR95: 68.45 -> 49.72Improved
add DICE weight sparsification to the energy scoreEnergy score · CIFAR-10 ID, avg of 6 OOD setsFPR95: 26.55 -> 20.83Improved
add DICE weight sparsification on top of ReActReAct · ImageNet-1k ID, avg of 4 OOD setsFPR95: 31.43 -> 27.25Improved
use directed (contribution-based) masking instead of random weight dropout (ablation)DICE · ImageNet-1k ID, avg of 4 OOD setsFPR95: 76.28 -> 34.75Improved
use contribution-directed masking instead of L1 weight pruning (ablation)DICE · ImageNet-1k ID, avg of 4 OOD setsFPR95: 52.81 -> 34.75Improved
use contribution-directed masking instead of random unit dropout (ablation)DICE · ImageNet-1k ID, avg of 4 OOD setsFPR95: 83.91 -> 34.75Improved
use contribution-directed masking instead of L2 unit pruning (ablation)DICE · ImageNet-1k ID, avg of 4 OOD setsFPR95: 90.80 -> 34.75Improved
keep top-k contributions instead of random-k (ablation)DICE · CIFAR-100 ID, avg of 6 OOD setsFPR95: 77.48 -> 49.72Improved
keep top-k contributions instead of bottom-k (ablation)DICE · CIFAR-100 ID, avg of 6 OOD setsFPR95: 99.70 -> 49.72Improved
keep top-k contributions instead of top+bottom-k (ablation)DICE · CIFAR-10 ID, avg of 6 OOD setsFPR95: 24.25 -> 20.83Improved
add mild sparsity p=0.1 versus none (p=0) (ablation)DICE · ImageNet-1k ID, avg of 4 OOD sets, sparsity sweepFPR95: 58.41 -> 40.25Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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