AI Paper Atlas

Research map / Continual learning / Class-incremental

CroMo-Mixup: Augmenting Cross-Model Representations for Continual Self-Supervised Learning

Erum Mushtaq, Duygu Nur Yaldiz, Yavuz Faruk Bakman, Jie Ding, Chenyang Tao, Dimitrios Dimitriadis, Salman Avestimehr

ECCV 2024 · Continual learning › Class-incremental · approach: Class-balanced knowledge retention · 4 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 cross-task data mixup and cross-model feature mixup to CSSLCaSSLe+ (Barlow-Twins) · CIFAR100-Split5average linear accuracy: 61.3 -> 65.5Improved
add CroMo-Mixup on CIFAR100-Split10Barlow-Twins CSSL · CIFAR100-Split10linear accuracy: 56.64 -> 62.48Improved
add CroMo-Mixup on CIFAR10-Split2 with Barlow-TwinsBarlow-Twins CSSL · CIFAR10-Split2Task-ID prediction: 95.87 -> 95.75No clear effect
split classes across tasks in minibatch training instead of uniform sampling (ablation)contrastive SSL baselines · CIFAR100 task confusion probelinear accuracy and task-ID prediction: drop of 4% or moreHurt
split classes across tasks in supervised training (ablation)supervised learning · CIFAR100 task confusion probelinear accuracyNo clear effect
split data (not classes) across tasks in SSL training (ablation)contrastive SSL baselines · CIFAR100 data-incremental probelinear accuracyNo clear effect
use cross-task instead of within-task data mixup (ablation)CroMo-Mixup · CIFAR100-Split5 (Barlow-Twins)accuracy: 54.16 -> 62.31Improved
use cross-model instead of same-model feature mixup (ablation)CroMo-Mixup · CIFAR100-Split5 (Barlow-Twins)accuracy: 62.31 -> 63.94Improved
add distillation with temporal projector (ablation)CroMo-Mixup · CIFAR100-Split5 (Barlow-Twins)accuracy: 63.94 -> 65.48Improved
reduce memory buffer to 25 samples per task (ablation)CroMo-Mixup · CIFAR100-Split10average linear accuracy: drops of 1-2 percentage pointsHurt
apply CroMo-Mixup to out-of-distribution transferResNet-18 trained on CIFAR100-Split5 · Flower102linear accuracy: 44.17 -> 49.03Improved

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