Views Can Be Deceiving: Improved SSL Through Feature Space Augmentation
원문 초록
Supervised learning methods have been found to exhibit inductive biases favoring simpler features. When such features are spuriously correlated with the label, this can result in suboptimal performance on minority subgroups. Despite the growing popularity of methods which learn from unlabeled data, the extent to which these representations rely on spurious features for prediction is unclear. In this work, we explore the impact of spurious features on Self-Supervised Learning (SSL) for visual representation learning. We first empirically show that commonly used augmentations in SSL can cause undesired invariances in the image space, and illustrate this with a simple example. We further show that classical approaches in combating spurious correlations, such as dataset re-sampling during SSL, do not consistently lead to invariant representations. Motivated by these findings, we propose LateTVG to remove spurious information from these representations during pre-training, by regularizing later layers of the encoder via pruning. We find that our method produces representations which outperform the baselines on several benchmarks, without the need for group or label information during SSL.
함께 살펴볼 논문
공통 태그를 근거로 연결했습니다. 실제 인용 관계나 논문의 우수성을 뜻하지 않습니다.
Self-supervised Learning of Echocardiographic Video Representations via Online Cluster Distillation
Subgraph Gaussian Embedding Contrast for Self-Supervised Graph Representation Learning
Advancing Video Self-Supervised Learning via Image Foundation Models
ShapeEmbed: a self-supervised learning framework for 2D contour quantification
USAD: Universal Speech and Audio Representation via Distillation
Exploring Non-contrastive Self-supervised Representation Learning for Image-based Profiling
토론
이 논문과 연결된 포럼 글이 아직 없습니다.