AI Paper Atlas

Research map / Continual learning / Online and replay-based

Lifelong Compression Mixture Model via Knowledge Relationship Graph

Fei Ye, Adrian G. Bors

AAAI 2023 · Continual learning › Online and replay-based · approach: Unified and unsupervised continual frameworks · 1 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 diversity-aware sample selection (DivSS) instead of sliding-window selection (ablation)LGMM · Split MNISTclassification accuracy: 96.81 ± 0.12 -> 96.95 ± 0.13Improved
Use diversity-aware sample selection (DivSS) instead of sliding-window selection (ablation)LGMM · Split CIFAR10classification accuracy: 50.91 ± 0.25 -> 53.71 ± 0.19Improved
Use diversity-aware sample selection (DivSS) instead of sliding-window selection (ablation)LGMM · Split CIFAR100classification accuracy: 25.65 ± 0.16 -> 26.03 ± 0.16Improved
Add component discarding to DivSS (ablation)LGMM · Split MNISTclassification accuracy: 96.95 ± 0.13 -> 96.16 ± 0.11Hurt
Add component discarding to DivSS (ablation)LGMM · Split CIFAR10classification accuracy: 53.71 ± 0.19 -> 50.12 ± 0.23Hurt
Add component discarding to DivSS (ablation)LGMM · Split CIFAR100classification accuracy: 26.03 ± 0.16 -> 25.24 ± 0.17Hurt
Add component discarding to DivSS (ablation)LGMM · Split MNIST-SVHNclassification accuracy: 65.29 -> 63.78Hurt
Add component discarding to DivSS (ablation)LGMM · Split MNIST-CIFAR10classification accuracy: 64.36 -> 62.15Hurt
Add component discarding to DivSS (ablation)LGMM · Split MNIST-SVHNnumber of components: 18 -> 13Improved
Add component discarding to DivSS (ablation)LGMM · Split MNIST-SVHNnumber of parameters: 175M -> 126MImproved
Add component discarding to DivSS (ablation)LGMM · Split MNIST-SVHNtesting time per sample: 0.0064 -> 0.0048Improved
Use a larger expansion threshold λ (ablation)LGMM · Split MNISTclassification accuracyHurt

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