Hugging Face's logo

Join the conversation

Join the community of Machine Learners and AI enthusiasts.

Sign Up
Back to feed
Kseniase 
posted an update May 25, 2025
Post
9351
14 Types of JEPA

JEPA, or Joint Embedding Predictive Architecture, is an approach to building AI models introduced by Yann LeCun. It differs from transformers by predicting the representation of a missing or future part of the input, rather than the next token or pixel. This encourages conceptual understanding, not just low-level pattern matching. So JEPA allows teaching AI to reason abstractly.

Here are 14 types of JEPA you should know about :)

▪️ JEPA / H-JEPA
▪️ I-JEPA
▪️ MC-JEPA
▪️ V-JEPA
▪️ Audio-JEPA
▪️ Point-JEPA
▪️ 3D-JEPA
▪️ ACT-JEPA
▪️ V-JEPA 2
▪️ LeJEPA
▪️ Causal-JEPA
▪️ V-JEPA 2.1
▪️ LeWorldModel
▪️ ThinkJEPA

Explore these JEPA milestones as a map of AI progress: https://www.turingpost.com/p/jepamap

Also, explore the basics of JEPA in our article: https://www.turingpost.com/p/jepa

If you liked it, subscribe to the Turing Post: https://www.turingpost.com/subscribe
Kseniase

  1. T-JEPA -> https://huggingface.co/papers/2410.05016
    This one is for tabular (structured) data. By masking one subset of a table’s features and predicting their latent representation from another subset, it learns rich, label-agnostic embeddings

  2. ACT-JEPA -> https://huggingface.co/papers/2501.14622
    Merges imitation and self-supervised learning to learn policy embeddings without heavy expert data. It predicts chunked actions and abstract observations in latent space, filtering noise, modeling dynamics, and cutting compounding errors

  3. Brain-JEPA -> https://huggingface.co/papers/2409.19407
    Applies JEPA in brain dynamics foundation model for demographic, disease, and trait prediction.

  4. 3D-JEPA -> https://huggingface.co/papers/2409.15803
    JEPA for 3D representation learning. It samples one rich context block and several target blocks, then predicts each target’s embedding from the context

  5. Point-JEPA -> https://huggingface.co/papers/2404.16432
    Brings joint-embedding predictive learning to point clouds. A lightweight sequencer orders patch embeddings. It lets the model choose context and target patches and reuse distance calculations for speed
In this post