AI Paper Atlas

Research map / Continual learning / Class-incremental

incDFM: Incremental Deep Feature Modeling for Continual Novelty Detection

Amanda Rios, Nilesh Ahuja, Ibrahima Ndiour, Utku Genc, Laurent Itti, Omesh Tickoo

ECCV 2022 · Continual learning › Class-incremental · approach: Class-balanced knowledge retention · 13 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 incremental recruitment of confident novel samples (incDFM vs single-threshold DFM) (ablation)DFM · CIFAR-10 class-incrementalAUROC (%): 74.0 -> 98.2Improved
add incremental recruitment of confident novel samples (ablation)DFM · CIFAR-10 class-incrementalAUPR (%): 72.1 -> 98.2Improved
add incremental recruitment of confident novel samples (ablation)DFM · EMNIST class-incrementalAUROC (%): 75.4 -> 98.7Improved
add incremental recruitment of confident novel samples (ablation)DFM · offline CIFAR-10 -> SVHNAUROC (%): 93.4 -> 99.9Improved
run two recruitment iterations instead of single thresholding (ablation)DFM · CIFAR-10AUROC: 22% improvementImproved
allow up to 10 maximum recruitment iterations (ablation)DFM · CIFAR-10AUROC: up to 39.4% improvementImproved
lower recruitment percentage per iteration (15%) (ablation)incDFM · CIFAR-10iterations to convergeHurt
use estimated novel samples instead of ground-truth novel samples for consolidation (ablation)incDFM · CIFAR-10AUROC (%): 98.2 -> 98.2No clear effect
use estimated novel samples instead of ground-truth novel samples for consolidation (ablation)ODIN · CIFAR-10AUROC (%): 87.3 -> 81.6Hurt
remove the task oracleSoftmax · 8-dataset continual novelty detectionAUROC (%): 99.9 -> 69.4Hurt
remove the task oracleODIN · 8-dataset continual novelty detectionAUROC (%): 99.8 -> 64.2Hurt
increase old:new imbalance from 1:1 to 1:4incDFM · CIFAR-10AUPR (%): 98.2 -> 95.0Hurt

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