Papers › Guarding Barlow Twins Against Overfitting with Mixed Samples
Guarding Barlow Twins Against Overfitting with Mixed Samples
Wele Gedara Chaminda Bandara, Celso M. de Melo, Vishal M. Patel
Self-supervised Learning (SSL) aims to learn transferable feature representations for downstream applications without relying on labeled data. The Barlow Twins algorithm, renowned for its widespread adoption and straightforward implementation compared to its counterparts like contrastive learning methods, minimizes feature redundancy while maximizing invariance to common corruptions. Optimizing for the above objective forces the network to learn useful representations, while avoiding noisy or constant features, resulting in improved downstream task performance with limited adaptation. Despite Barlow Twins' proven effectiveness in pre-training, the underlying SSL objective can inadvertently cause feature overfitting due to the lack of strong interaction between the samples unlike the contrastive learning approaches. From our experiments, we observe that optimizing for the Barlow Twins objective doesn't necessarily guarantee sustained improvements in representation quality beyond a certain pre-training phase, and can potentially degrade downstream performance on some datasets. To address this challenge, we introduce Mixed Barlow Twins, which aims to improve sample interaction during Barlow Twins training via linearly interpolated samples. This results in an additional regularization term to the original Barlow Twins objective, assuming linear interpolation in the input space translates to linearly interpolated features in the feature space. Pre-training with this regularization effectively mitigates feature overfitting and further enhances the downstream performance on CIFAR-10, CIFAR-100, TinyImageNet, STL-10, and ImageNet datasets. The code and checkpoints are available at: https://github.com/wgcban/mix-bt.git
Code
Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.
Code Syntology ran Syntology
Not run by Syntology. Nothing on this page verifies that the listed code works.
Tasks
Results from the paper archive 2025-07-28
| Task | Dataset | Model | Metric | Value | Rank at snapshot | Leaderboard | Report |
|---|---|---|---|---|---|---|---|
| Self-Supervised Learning | STL-10 | ResNet50 | Accuracy | 91.70 | #1 of 3 | Archive leaderboard | report |
| Self-Supervised Learning | STL-10 | ResNet18 | Accuracy | 91.02 | #2 of 3 | Archive leaderboard | report |
| Self-Supervised Learning | TinyImageNet | ResNet18 | average top-1 classification accuracy | 51.67 | #1 of 2 | Archive leaderboard | report |
| Self-Supervised Learning | TinyImageNet | ResNet50 | average top-1 classification accuracy | 51.84 | #2 of 2 | Archive leaderboard | report |
| Self-Supervised Learning | cifar10 | ResNet50 | average top-1 classification accuracy | 93.89 | #1 of 2 | Archive leaderboard | report |
| Self-Supervised Learning | cifar10 | ResNet18 | average top-1 classification accuracy | 92.58 | #2 of 2 | Archive leaderboard | report |
| Self-Supervised Learning | cifar100 | ResNet50 | average top-1 classification accuracy | 72.51 | #1 of 2 | Archive leaderboard | report |
| Self-Supervised Learning | cifar100 | ResNet18 | average top-1 classification accuracy | 69.31 | #2 of 2 | Archive leaderboard | report |
Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.
Methods
Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections