# PEIRA: Learning Predictive Encoders through Inter-View Regressor Alignment

Source: [PEIRA: Learning Predictive Encoders through Inter-View Regressor Alignment](<https://www.sota2.com/research/paper/6a0beeeadbc836b87d82c859>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-05-20T05:00:13.583Z

Authors: Michael Arbel, Basile Terver, Jean Ponce

Published: 2026-05-17T22:04:01.000Z

## 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 Architecture (JEPA) using a regularized linear regressor to predict the learned representations of two views of the data from one another, and fully characterize its stability: non-collapsed stable equilibria align with leading nonlinear canonical correlation subspaces, while collapsed equilibria may also be stable attractors. Motivated by this result, we introduce PEIRA, a non-contrastive SSL method with an explicit objective defined through the trace of the optimal linear regressor. We show that its only stable equilibria are nontrivial global minimizers and recover the same canonical correlation subspaces, with regularization selecting the effective dimension. Experiments on ImageNet-1K and CIFAR-10 show PEIRA is competitive with VICReg and LeJEPA baselines, and qualitative empirical results support the theory.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Classification on CIFAR-10 (train test)](<https://www.sota2.com/research/sota/classification-on-cifar-10-train-test.md>): Classification; CIFAR-10 (train/test); Top-1 Accuracy: 90.97.

- [Image Classification on ImageNet 1K (train val)](<https://www.sota2.com/research/sota/image-classification-on-imagenet-1k-train-val.md>): Image Classification; ImageNet 1K (train val); Top-1 Accuracy: 66.5.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2605.17671>)

[PDF](<https://arxiv.org/pdf/2605.17671.pdf>)
