AI Paper Atlas

Research map / Continual learning / Continual fine-tuning of LLMs

Continual Learners are Incremental Model Generalizers

Jaehong Yoon, Sung Ju Hwang, Yue Cao

ICML 2023 · Continual learning › Continual fine-tuning of LLMs · approach: Low-rank and model-merging adaptation · 1 citations (OpenAlex)

Read the paper

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 masked modeling instead of supervised continual learningBase continual learner · CIFAR-100 Split, fine-tuningTop-1 accuracy: 79.3 -> 88.9Improved
Add GLAD attention adaptors and regularizationSupervised continual pre-training · ImageNet-1K Split, T9 fine-tuning after T5-T8accuracy: 64.56 -> 65.78; 69.06 -> 69.84; 75.06 -> 76.12; 77.80 -> 79.04Improved
Add GLAD attention adaptors and regularizationSimMIM continual pre-training · ImageNet-1K Split, T9 fine-tuning after T5-T8accuracy: 68.34 -> 68.24; 72.22 -> 73.24; 77.38 -> 78.94; 80.12 -> 81.62Mixed
Add GLAD attention adaptors and regularizationMAE continual pre-training · ImageNet-1K Split, T9 fine-tuning after T5-T8accuracy: 42.19 -> 41.75; 52.07 -> 54.74; 63.07 -> 68.01; 71.19 -> 74.30Mixed
Freeze the two lowest or two deepest layers during subsequent continual learning (ablation)Supervised continual learning · ImageNet-1K Split, unseen-task fine-tuningaccuracyHurt
Freeze the two lowest or two deepest layers during subsequent continual learning (ablation)MIM-based unsupervised continual learning · ImageNet-1K Split, unseen-task fine-tuningaccuracyNo clear effect
Train on additional sequential tasksContinual learners · ImageNet-1K Split, unseen-task T9fine-tuning 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