2025/04/06 by Mehmet Can Yavuz, Yavuz, Mehmet Can, Berrin Yanıkoğlu +1
Computer Science · #Bridging (networking) #Computer Vision and Pattern Recognition (cs.CV) #Education and Learning Interventions #Encoder #FOS: Computer and information sciences #Feature learning #Gaussian #Gaussian process #Generative grammar #Generative model #Inference #Machine Learning (cs.LG) #Representation (politics)
paper · pdf · doi:10.48550/arxiv.2504.04318
openalex publication_date 2025/04/06 · openalex created_date 2025/10/11 · openalex updated_date 2026/08/05
We present Variational Self-Supervised Learning (VSSL), a novel framework that combines variational inference with self-supervised learning to enable efficient, decoder-free representation learning. Unlike traditional VAEs that rely on input reconstruction via a decoder, VSSL symmetrically couples two encoders with Gaussian outputs. A momentum-updated teacher network defines a dynamic, data-dependent prior, while the student encoder produces an approximate posterior from augmented views. The reconstruction term in the ELBO is replaced with a cross-view denoising objective, preserving the analytical tractability of Gaussian KL divergence. We further introduce cosine-based formulations of KL and log-likelihood terms to enhance semantic alignment in high-dimensional latent spaces. Experiments on CIFAR-10, CIFAR-100, and ImageNet-100 show that VSSL achieves competitive or superior performance to leading self-supervised methods, including BYOL and MoCo V3. VSSL offers a scalable, probabilistically grounded approach to learning transferable representations without generative reconstruction, bridging the gap between variational modeling and modern self-supervised techniques.