MC-JEPA: A Joint-Embedding Predictive Architecture for Self-Supervised Learning of Motion and Content Features — mentioned by jamesblonde

Mon, 25 Sep 2023 04:59:34 GMT

You are correct that we have two modalities projecting into a shared latent space. However, that doesn't convey anything about how to jointly train the embedding models, does it? That's what the two tower embedding model shows - how to jointly train the models using distance functions like cosine distance. On an aside, others, as esteemed as Yann Lecun, are talking about learning in a shared latent space between 2 modalities, but use different names - joint embeddings, because it's a different method. https://arxiv.org/abs/2307.12698

Paper and discussion history. A mention is not an endorsement.

MC-JEPA: A Joint-Embedding Predictive Architecture for Self-Supervised Learning of Motion and Content Features — HN submission

Thu, 27 Jul 2023 21:22:45 GMT

MC-JEPA

1 point; 1 discussion reply on HN. Paper and source history. A submission is not an endorsement.

MC-JEPA: A Joint-Embedding Predictive Architecture for Self-Supervised Learning of Motion and Content Features · HackerPapers discussions

MC-JEPA: A Joint-Embedding Predictive Architecture for Self-Supervised Learning of Motion and Content Features — mentioned by jamesblonde

MC-JEPA: A Joint-Embedding Predictive Architecture for Self-Supervised Learning of Motion and Content Features — HN submission