AI Paper Atlas

Research map / Continual learning / Online and replay-based

Overcoming Catastrophic Forgetting by Neuron-Level Plasticity Control

Inyoung Paik, Sangjun Oh, Taeyeong Kwak, Injung Kim

AAAI 2020 · Continual learning › Online and replay-based · approach: Gradient projection and plasticity preservation · 50 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 neuron-level plasticity control instead of MAS connection-level consolidationCNN · iMNISTaverage validation accuracy: 98.76(0.38) -> 99.63(0.02)Improved
Use neuron-level plasticity control instead of MAS connection-level consolidationCNN · iCIFAR100average validation accuracy: 65.54(0.73) -> 71.87(0.65)Improved
Control plasticity at the neuron level rather than the connection level (ablation)NPC on CNN · iMNISTaverage validation accuracy: 96.49(0.65) -> 99.63(0.02)Improved
Control plasticity at the neuron level rather than the connection level (ablation)NPC on CNN · iCIFAR100average validation accuracy: 50.70(1.27) -> 71.87(0.65)Improved
Add NPC to protect neurons according to their importanceCNN second-top-layer neurons · iCIFAR100 Task 1 neurons after learning Task 2average activation change of most important 10% of neurons: 0.383 -> 0.094Improved
Add NPC to allow less important neurons greater plasticityCNN second-top-layer neurons · iCIFAR100 Task 1 neurons after learning Task 2average activation change of least important 10% of neurons: 0.383 -> 0.667Mixed

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