AI Paper Atlas

Research map / Continual learning / Domain and prompt-based

When Prompt-based Incremental Learning Does Not Meet Strong Pretraining

Yu-Ming Tang, Yi-Xing Peng, Wei-Shi Zheng

ICCV 2023 · Continual learning › Domain and prompt-based · approach: Prompt-based continual learning · 42 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
add L_conC knowledge-pool constraint on the classifier (ablation)APG · ImageNet-Sub B50-T10Avg. Acc.: 19.37 -> 71.84Improved
add L_conA constraint on the APG (ablation)APG · ImageNet-Sub B50-T10Avg. Acc.: 71.84 -> 73.46Improved
add class-specific attention loss L_attn (ablation)APG · ImageNet-Sub B50-T10Avg. Acc.: 73.46 -> 74.70Improved
add triplet loss L_tri (ablation)APG · ImageNet-Sub B50-T10Avg. Acc.: 74.70 -> 75.94Improved
place APG at layer 10 instead of layers 6-9 (ablation)APG · ImageNet-Sub B50-T10Avg. Acc.Hurt
generate 1 prompt instead of none (N_P=0) (ablation)APG · ImageNet-Sub B50-T10Avg. Acc.: +4.1%Improved
increase number of generated prompts beyond 1 (ablation)APG · ImageNet-Sub B50-T10Avg. Acc.No clear effect
replace prompt-pool retrieval with learnable adaptive prompt generator (non-pretrained)L2P / DualPrompt · ImageNet-Subset B50-T10Avg. Acc.: 35.82 -> 75.52 (vs DualPrompt)Improved
use APG with Tiny-ImageNet pretraining instead of prompt poolL2P / DualPrompt · CIFAR100 B0-20 (Tiny-ImageNet pretrained)Avg. Acc.: 79.13 -> 87.50 (vs DualPrompt)Improved
use APG with strong ImageNet-21k pretraining, 5 tasksDualPrompt · CIFAR100 B0-5 (ImageNet-21K pretrained)Avg. Acc.: 89.56 -> 88.48No clear effect
use APG instead of DualPrompt with 20 tasks under ImageNet-21k pretrainingDualPrompt · ImageNet-R B0-T20Avg. Acc.: 70.48 -> 71.22Improved
exemplar-free APG vs rehearsal methods with 20 exemplars/class, 50 tasksAFC · ImageNet-Subset B50-T50Avg. Acc.: 71.97 -> 75.70Improved

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