Drive vs. Decay: On the Training Dynamics of Joint-Embedding Predictive Architectures
cs.LG
Submitted: 2026-10-01
Updated: 2026-10-01
Code: https://github.com/jose-melo/drive-vs-decay
Sources
- LeJEPA: Provable and Scalable Self-Supervised Learning Without the Heuristics
- Video Representation Learning with Joint-Embedding Predictive Architectures
- A-JEPA: Joint-Embedding Predictive Architecture Can Listen
- Why and How Auxiliary Tasks Improve JEPA Representations
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks