AI Paper Atlas

Research map / Continual learning / Online and replay-based

Meta-Consolidation for Continual Learning

Joseph K J, Vineeth N Balasubramanian

NeurIPS 2020 · Continual learning › Online and replay-based · approach: Unified and unsupervised continual frameworks · 34 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
train VAE meta-distribution and consolidate (vs skipping) (ablation)MERLIN · Split CIFAR-100average accuracy: 15.7 -> 43.5Improved
train VAE meta-distribution and consolidate (vs skipping) (ablation)MERLIN · Permuted MNISTaverage accuracy: 29.9 -> 85.5Improved
ensemble 30 sampled models instead of 1 (ablation)MERLIN · Split Mini-ImageNetaverage accuracy: 30.1 -> 39.1Improved
ensemble 100 instead of 30 models (ablation)MERLIN · Split MNISTaverage accuracy: 90.6 -> 91.6No clear effect
increase chunk size from 300 to 2000 (ablation)MERLIN · Split Mini-ImageNetaverage accuracy: 40.5 -> 37.8Hurt
task-aware instead of task-agnostic inference (ablation)MERLIN · Split MNISTaverage accuracy: 90.7 -> 97.4Improved
task-aware instead of task-agnostic inference (ablation)MERLIN · Split CIFAR-10average accuracy: 82.9 -> 82.3No clear effect
increase exemplar memory from 100 to 2000 (ablation)MERLIN · Split CIFAR-10average accuracy: 77.9 -> 88.4Improved
add auxiliary classification loss to ELBO (ablation)MERLIN · Split Mini-ImageNetaverage accuracy: 40.1 -> 40.8No clear effect
add auxiliary classification loss to ELBO (ablation)MERLIN · Split CIFAR-100average accuracy: 43.5 -> 42.5Mixed
use MERLIN meta-consolidation instead of GEM/iCaRL/GSS (ablation)MERLIN · Split CIFAR-100average accuracy: 40.6 -> 43.5Improved
MERLIN vs iCaRL on forgettingMERLIN · Split MNISTforgetting: 1.7 -> 6.4Hurt

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