AI Paper Atlas

Research map / Continual learning / Online and replay-based

SparCL: Sparse Continual Learning on the Edge

Zifeng Wang, Zheng Zhan, Yifan Gong, Geng Yuan, Wei Niu, Tong Jian, Bin Ren, Stratis Ioannidis, Yanzhi Wang, Jennifer Dy

NeurIPS 2022 · Continual learning › Online and replay-based · approach: Unified and unsupervised continual frameworks · 15 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
Apply SparCL with weight sparsity 0.75 to DER++DER++ · Split CIFAR-10 Class-IL, buffer 500average accuracy: 72.70 -> 74.09Improved
Apply SparCL with weight sparsity 0.75 to DER++DER++ · Split Tiny-ImageNet Class-IL, buffer 500average accuracy: 19.38 -> 20.75Improved
Use SparCL weight sparsity 0.95 with DER++DER++ · Split CIFAR-10 Class-IL, buffer 500training FLOPs (× 10^15): 13.9 -> 0.6Improved
Use SparCL instead of SNIP sparse training with DER++DER++ · Split CIFAR-10 Class-IL, weight sparsity 0.90average accuracy: 66.07 -> 73.42Improved
Use SparCL instead of RigL sparse training with DER++DER++ · Split Tiny-ImageNet Class-IL, weight sparsity 0.90average accuracy: 15.88 -> 19.62Improved
Add task-aware dynamic masking (TDM) (ablation)SparCL-DER++ · Split CIFAR-10 Class-IL, weight sparsity 0.75average accuracy: 72.70 -> 73.37Improved
Add task-aware dynamic masking (TDM) (ablation)SparCL-DER++ · Split CIFAR-10 Class-IL, weight sparsity 0.75training FLOPs (× 10^15): 13.9 -> 3.6Improved
Add dynamic data removal (DDR) after TDM (ablation)SparCL-DER++ · Split CIFAR-10 Class-IL, weight sparsity 0.75average accuracy: 73.37 -> 73.80Improved
Add dynamic data removal (DDR) after TDM (ablation)SparCL-DER++ · Split CIFAR-10 Class-IL, weight sparsity 0.75training FLOPs (× 10^15): 3.6 -> 2.8Improved
Add dynamic gradient masking (DGM) after TDM (ablation)SparCL-DER++ · Split CIFAR-10 Class-IL, weight sparsity 0.75average accuracy: 73.37 -> 73.97Improved
Add dynamic gradient masking (DGM) after TDM (ablation)SparCL-DER++ · Split CIFAR-10 Class-IL, weight sparsity 0.75training FLOPs (× 10^15): 3.6 -> 3.3Improved
Increase data removal proportion from 30% to at least 40% (ablation)DDR · Split CIFAR-10 Class-ILaverage 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