cs.CVNov 17, 2026

CascadeJEPA: Enhancing 3D Geometric Representations via Spatially Coupled Cascaded Prediction

Authors: Kailai Zhou, Ziqi Chen, Xiaoning Xu, Yuefeng Wu, Zixuan Fu, Zijun Lin, Zhihao Li, Youwei Pang, 

Organizations: Nanyang Technological University · Nanjing University

Abstract

Joint-Embedding Predictive Architectures (JEPAs) have emerged as a promising paradigm for predictive visual representation learning. However, our analysis of I-JEPA reveals a growing discrepancy between semantic and geometric learning: improved semantic recognition during pretraining is accompanied by degraded 3D geometric representations. We hypothesize that this partly stems from predicting highly abstract latent representations with incomplete spatial context, favoring semantic regularities over fine-grained geometric information. To address this issue, we introduce CascadeJEPA, a spatially coupled cascaded prediction framework that promotes richer patch-level interactions through full-grid context encoding and cross-attention between prediction stages, while largely preserving the original I-JEPA architecture. Extensive experiments demonstrate substantial improvements over I-JEPA across multiple 3D geometric tasks while maintaining strong semantic recognition performance. Furthermore, its advantages extend to future feature forecasting and world model planning, highlighting the potential of enhanced geometric representations to support predictive world modeling.

Figures & tables

Explore similar work

CardsList
  1. SR-JEPA: Learning Predictive Latent State in 3D Scenes

    Aug 6, 2026Zihan Zhou, Qifu Wen, Xi ZengJoint-Embedding Predictive ArchitectureObject-Centric Representation Learning

  2. Slot-JEPA: Object-Structured Prediction Interfaces for Latent World Models

    Oct 20, 2026Peisong Wen, Qianqian Xu, Minye Lei +3Goal-Conditioned PlanningJoint-Embedding Predictive Architecture