HP-JEPA: Hierarchical Partitioning for Multi-Resolution Graph Joint-Embedding Predictive Learning
Published in arXiv preprint, 2026
HP-JEPA learns from a hierarchy of graph partitions, predicting target embeddings at each resolution. Combining the resulting representations lets downstream tasks use complementary structural information at multiple scales.
Recommended citation: Ruichen Xu, Jingxiang Qu, Wenhan Gao, Jiaxing Zhang, Linsey Pang, Ravid Shwartz-Ziv, Yann LeCun, Yuefan Deng. 2026. HP-JEPA: Hierarchical Partitioning for Multi-Resolution Graph Joint-Embedding Predictive Learning. arXiv:2608.00491.
Recommended citation: Ruichen Xu, Jingxiang Qu, Wenhan Gao, Jiaxing Zhang, Linsey Pang, Ravid Shwartz-Ziv, Yann LeCun, Yuefan Deng. 2026. HP-JEPA: Hierarchical Partitioning for Multi-Resolution Graph Joint-Embedding Predictive Learning. arXiv:2608.00491. https://arxiv.org/abs/2608.00491