mirror of
https://github.com/hanzoai/papers.git
synced 2026-10-30 14:58:42 +00:00
35 lines
1.4 KiB
TeX
35 lines
1.4 KiB
TeX
\section{Conclusion}
|
|
|
|
|
|
Jin demonstrates that joint embedding architectures can unify multiple modalities into a single, coherent representation space. The combination of modality-specific encoders, diffusion transformer backbone with MoE, and hierarchical z-JEPA enables state-of-the-art performance on multimodal understanding tasks while maintaining computational efficiency. The architecture's extensibility opens new possibilities for AI systems that perceive and interact with the world across all sensory modalities.
|
|
|
|
\bibliographystyle{plain}
|
|
\begin{thebibliography}{10}
|
|
|
|
\bibitem{jepa2022}
|
|
Y. LeCun. A Path Towards Autonomous Machine Intelligence. 2022.
|
|
|
|
\bibitem{moe2017}
|
|
N. Shazeer et al. Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer. ICLR, 2017.
|
|
|
|
\bibitem{ddpm2020}
|
|
J. Ho et al. Denoising Diffusion Probabilistic Models. NeurIPS, 2020.
|
|
|
|
\bibitem{dit2023}
|
|
W. Peebles, S. Xie. Scalable Diffusion Models with Transformers. ICCV, 2023.
|
|
|
|
\bibitem{vit2020}
|
|
A. Dosovitskiy et al. An Image is Worth 16x16 Words. ICLR, 2021.
|
|
|
|
\bibitem{ijepa2023}
|
|
M. Assran et al. Self-Supervised Learning from Images with a Joint-Embedding Predictive Architecture. CVPR, 2023.
|
|
|
|
\bibitem{vjepa2024}
|
|
A. Bardes et al. V-JEPA: Latent Video Prediction for Visual Representation Learning. 2024.
|
|
|
|
\bibitem{clip2021}
|
|
A. Radford et al. Learning Transferable Visual Models From Natural Language Supervision. ICML, 2021.
|
|
|
|
\end{thebibliography}
|
|
|