Joint Embedding vs Reconstruction: Provable Benefits of Latent Space Prediction for Self Supervised Learning
Paper • 2505.12477 • Published • 1
Brown University
university
Verified
AI & ML interests
None defined yet.
Recent Activity
Papers
Deform360: A Massive Multi-view Visuotactile Dataset for Deformable World Models
A Cookbook of 3D Vision: Data, Learning Paradigms, and Application
RandallBalestriero
authored 20
papers 5 days ago
GrokAlign: Geometric Characterisation and Acceleration of Grokking
Paper • 2506.12284 • Published
FastDINOv2: Frequency Based Curriculum Learning Improves Robustness and Training Speed
Paper • 2507.03779 • Published • 1
Deep Networks Always Grok and Here is Why
Paper • 2402.15555 • Published
LeJEPA: Provable and Scalable Self-Supervised Learning Without the Heuristics
Paper • 2511.08544 • Published • 16
Max-Affine Spline Insights Into Deep Network Pruning
Paper • 2101.02338 • Published
stable-pretraining-v1: Foundation Model Research Made Simple
Paper • 2511.19484 • Published • 1
DIET-CP: Lightweight and Data Efficient Self Supervised Continued Pretraining
Paper • 2509.06990 • Published • 1
LLM-JEPA: Large Language Models Meet Joint Embedding Predictive Architectures
Paper • 2509.14252 • Published • 10
Gaussian Embeddings: How JEPAs Secretly Learn Your Data Density
Paper • 2510.05949 • Published • 3
Temporal Straightening for Latent Planning
Paper • 2603.12231 • Published • 3
LeWorldModel: Stable End-to-End Joint-Embedding Predictive Architecture from Pixels
Paper • 2603.19312 • Published • 58
stable-worldmodel: A Platform for Reproducible World Modeling Research and Evaluation
Paper • 2605.21800 • Published
VideoGPA: Distilling Geometry Priors for 3D-Consistent Video Generation
Paper • 2601.23286 • Published • 1
When Does LeJEPA Learn a World Model?
Paper • 2605.26379 • Published
Rectified LpJEPA: Joint-Embedding Predictive Architectures with Sparse and Maximum-Entropy Representations
Paper • 2602.01456 • Published • 1
Hierarchical Planning with Latent World Models
Paper • 2604.03208 • Published • 4
Prompt Augmentation Scales up GRPO Training on Mathematical Reasoning
Paper • 2602.03190 • Published • 1
A Lightweight Library for Energy-Based Joint-Embedding Predictive Architectures
Paper • 2602.03604 • Published
VISReg: Variance-Invariance-Sketching Regularization for JEPA training
Paper • 2606.02572 • Published • 4