WeWorkBuddy

PEIRA: Learning Predictive Encoders through Inter-View Regressor Alignment

arXiv:2605.17671v2 Announce Type: replace-cross Abstract: Non-contrastive self-supervised learning (SSL) is an effective framework for predictive representation learning, but popular (and in practice effective) methods such as SimSiam, BYOL, I-JEPA or DINO, which rely on a form of self-distillation to train a teacher-student network, remain poorly understood as they typically do not minimize a well-defined objective. We analyze the dynamics of a variant of the Joint Embedding Predictive Architec

Read original source
Source reference for PEIRA: Learning Predictive Encoders through Inter-View Regressor Alignment. Screenshot not yet available.
Source record · preview pending · arxiv.orgOpen the source for the full page.

Source record

Project analysis

Project analysis is not yet available. See the source record above.