AI Paper Atlas

Research map / Continual learning / Online and replay-based

Continual Deep Learning by Functional Regularisation of Memorable Past

Pingbo Pan, Siddharth Swaroop, Alexander Immer, Runa Eschenhagen, Richard Turner, Mohammad Emtiyaz Khan

NeurIPS 2020 · Continual learning › Online and replay-based · approach: Theoretical analysis of forgetting · 54 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 FROMP instead of EWCcontinual learning · Split CIFARforward transfer: 6.1±0.7% vs 0.17±0.9%Improved
Use FROMP instead of VCL+coresetscontinual learning · Split CIFARforward transfer: 6.1±0.7% vs 1.8±3.1%Improved
Use FROMP instead of VCL+coresetscontinual learning · Split CIFARbackward transfer: -2.6±0.9% vs -9.2±1.8%Improved
Use FROMP instead of EWCcontinual learning · Split CIFARbackward transfer: -2.6±0.9% vs -2.3±1.4%No clear effect
Select memorable examples by the loss Hessian relevance instead of randomly (ablation)FROMP · Split CIFAR with 10 memorable examplesaverage accuracy: 45% -> 70%Improved
Select memorable examples by the loss Hessian relevance instead of randomly (ablation)FROMP · Split MNISTaverage accuracyNo clear effect
Replace the kernel matrix with the identity matrix (ablation)FROMP · Permuted and Split MNISTaverage accuracyHurt
Replace the kernel matrix with the identity matrix (ablation)FROMP · Split CIFARaverage accuracyNo clear effect
Use random memorable examples instead of selected memorable examples (ablation)FROMP · Split CIFAR with small memoryaverage accuracy: 70% vs 45%Hurt
Use FROMP instead of separate per-task trainingcontinual learning · Split CIFARaverage accuracyImproved
Use FROMP instead of joint trainingcontinual learning · Split CIFAR tasks 4–6accuracyNo clear effect
Use FROMP instead of weight-regularisation and FRCL methodscontinual learning · Permuted MNIST and Split MNISTaverage accuracyImproved

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