Official code for the paper Graph-level Representation Learning with Joint-Embedding Predictive Architectures published in Transactions on Machine Learning Research (TMLR)

Graph-JEPA: GNN+LLM with JEPA pretraining

No description available

graph-based joint embedding predictive architecture

This is a project of machine learning course at HCMUS

No description available

No description available

Neural World Model (NWM) is a cognitive world modeling system that goes beyond latent only predictive models like Dreamer, JEPA, or RSSM by explicitly representing entities, relations, events, intuition, memory, goals, and plans in a unified probabilistic world graph.