paper
AdaJEPA: An Adaptive Latent World Model
paperactiveprovisional
adajepa-an-adaptive-latent-world-model-69777f4a·1 events·first seen 2d agoAliases: AdaJEPA: An Adaptive Latent World Model
Co-occurring entities
More like this (12)
Joint Embedding Predictive Architecture (JEPA)Latent Context Language ModelsLatent World Recovery for Multimodal Learning with Missing ModalitiesAdaJEPALatent World Recoverylatent dynamical systemsLatent Consistency ModelsSIMMER: Benchmarking Latent Failures in LLM Executable Planning with a World ModelJoint Energy-Based Models (JEMs)Latent Profile AnalysisAdaptive Multi-Resolution Procedural Knowledge Compression for Large Language ModelsLatentMoE
Recent events (1)
AdaJEPA: Adaptive Latent World Model with Test-Time Self-Supervised Recalibration
AdaJEPA is a latent world model that performs test-time adaptation within model predictive control (MPC) loops, addressing the common failure mode of frozen world models under distribution shift. After training, the model uses observed state transitions as self-supervised adaptation signals to continuously recalibrate predictions during planning, requiring as few as one gradient step per replanning step. Evaluated on goal-reaching tasks, AdaJEPA substantially improves planning success rates compared to static world models.