AI Paper Atlas

Research map / Continual learning / Online and replay-based

BNS: Building Network Structures Dynamically for Continual Learning

Qi Qin, Wenpeng Hu, Han Peng, Dongyan Zhao, Bing Liu

NeurIPS 2021 · Continual learning › Online and replay-based · approach: Theoretical analysis of forgetting · 17 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 all three task similarity measures (cos, KL, Lp) as agent input vs none (ablation)BNS · CIFAR100 A50final ACC: 78.50 -> 82.39Improved
add all three task similarity measures vs none (ablation)BNS · F-EMNISTfinal ACC: 57.75 -> 60.38Improved
add all three task similarity measures vs none (ablation)BNS · F-Celebafinal ACC: 62.28 -> 65.44Improved
add all three task similarity measures vs none (ablation)BNS · MNISTfinal ACC: 99.24 -> 99.87Improved
add all three task similarity measures vs none (ablation)BNS · CIFAR10final ACC: 88.99 -> 91.40Improved
use cos+KL+Lp similarity instead of cosine only (ablation)BNS · CIFAR100 A50final ACC: 80.21 -> 82.39Improved
use cos+KL+Lp similarity instead of cosine only (ablation)BNS · F-Celebafinal ACC: 66.62 -> 65.44Hurt
use cos+KL+Lp similarity instead of cosine only (ablation)BNS · MNISTfinal ACC: 99.85 -> 99.87No clear effect
replace HAT's fixed masking with RL-built per-task network structure and initializationHAT · F-Celebafinal ACC: 59.82 -> 65.44Improved
RL-built structure and similarity-driven initialization instead of HATHAT · MNIST A5final ACC: 99.74 -> 99.87No clear effect
BNS vs HAT on CIFAR100 A50HAT · CIFAR100 A50BWT: 0.00 -> -4.35Hurt

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