DSeq-JEPA: Discriminative Sequential Joint-Embedding Predictive Architecture

TL;DR

DSeq-JEPA introduces a sequential, discriminative approach to self-supervised visual representation learning, inspired by human visual perception, improving the transferability and discriminative power of learned features across various vision tasks.

Contribution

It proposes a novel sequential, discriminative learning architecture that enhances self-supervised visual representations by incorporating attention-based region prediction order.

Findings

Outperforms I-JEPA variants on multiple vision tasks

Learns more discriminative and generalizable features

Effective across classification, detection, segmentation, and reasoning tasks

Abstract

Recent advances in self-supervised visual representation learning have demonstrated the effectiveness of predictive latent-space objectives for learning transferable features. In particular, Image-based Joint-Embedding Predictive Architecture (I-JEPA) learns representations by predicting latent embeddings of masked target regions from visible context. However, it predicts target regions in parallel and all at once, lacking ability to order predictions meaningfully. Inspired by human visual perception, which attends selectively and progressively from primary to secondary cues, we propose DSeq-JEPA, a Discriminative Sequential Joint-Embedding Predictive Architecture that bridges latent predictive and autoregressive self-supervised learning. Specifically, DSeq-JEPA integrates a discriminatively ordered sequential process with JEPA-style learning objective. This is achieved by (i) identifying primary discriminative regions using an attention-derived saliency map that serves as a proxy for visual importance, and (ii) predicting subsequent regions in discriminative order, inducing a curriculum-like semantic progression from primary to secondary cues in pre-training. Extensive experiments across tasks -- image classification (ImageNet), fine-grained visual categorization (iNaturalist21, CUB, Stanford Cars), detection/segmentation (MS-COCO, ADE20K), and low-level reasoning (CLEVR) -- show that DSeq-JEPA consistently learns more discriminative and generalizable representations compared to I-JEPA variants. Project page: https://github.com/SkyShunsuke/DSeq-JEPA.