AI Paper Atlas

Research map / Continual learning / Online and replay-based

Online Continual Learning under Extreme Memory Constraints

Enrico Fini, Stéphane Lathuilière, Enver Sangineto, Moin Nabi, Elisa Ricci

ECCV 2020 · Continual learning › Online and replay-based · approach: Plug-and-play online training enhancements · 59 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 dynamic gradient balancing of distillation and new-task gradients (ablation)BLD · CIFAR10 (2 tasks)average accuracy: 73.8 -> 78.0Improved
add dynamic gradient balancing of distillation and new-task gradients (ablation)BLD · CIFAR10 (5 tasks)average accuracy: 81.6 -> 86.2Improved
use joint training stage instead of alternating learning and distillation steps (ablation)BLD · CIFAR10 (2 tasks)average accuracy: 71.8 -> 78.0Improved
use joint training stage instead of alternating learning and distillation steps (ablation)BLD · CIFAR10 (5 tasks)average accuracy: 81.3 -> 86.2Improved
add two-stage distillation without balancing (ablation)BLD · CIFAR10 (5 tasks)average accuracy: 72.3 -> 81.6Improved
add batch-level distillation (BLD) to FinetuneFinetune · CIFAR10 (5 tasks)average accuracy: 72.3 -> 86.2Improved
add batch-level distillation (BLD) to FinetuneFinetune · MNIST (5 tasks)average accuracy: 80.6 -> 89.9Improved
add batch-level L2 parameter regularization to FinetuneFinetune · MNIST (5 tasks)average accuracy: 80.6 -> 78.0Hurt
replace batch-level L2 with BLDL2 · MNIST (2 tasks)average accuracy: 95.7 -> 94.5Hurt
compare BLD with single-pass LwF that needs stored task dataSingle-pass LwF · CIFAR10 (5 tasks)average accuracy: 86.2 -> 86.2No clear effect
use BLD instead of offline multi-epoch LwFOffline LwF · CIFAR10 (5 tasks)average accuracy: 94.8 -> 86.2Hurt
iterate the joint training stage twice (ablation)BLD · MC-OCLaccuracyImproved

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