The paper extends the analysis of Joint-Embedding Predictive Architectures (JEPAs) beyond Euclidean latent spaces to Riemannian manifolds. It shows that when latent variables lie on a sphere and the target distribution matches this spherical geometry, every optimal representation recovers the latent state up to an orthogonal transformation, demonstrating that Gaussian uniqueness is not universal. Experiments confirm that geometrically compatible targets improve linear recovery, especially in high-dimensional toroidal settings.
arXiv:2608.28840v1 Announce Type: new Abstract: Independently trained neural networks tend to encode the same data with similar latent geometries. These latent geometries are not directly compatible,...
Joint-embedding predictive architectures (JEPAs) learn latent dynamics for planning and avoid representation collapse by matching features to maximum-entropy distributions such as isotropic Gaussians,...
arXiv:2608.29029v3 Announce Type: replace-cross Abstract: Joint-Embedding Predictive Architectures (JEPAs) provide a powerful framework for latent world modeling and planning in a reconstruction-free...
arXiv:2608.29434v1 Announce Type: cross Abstract: JEPA world models make latent-space planning a practical route to control, but they are built almost exclusively on images. Whether latent prediction...
The paper introduces ATLAS, a training objective that preserves relational geometry in latent world models while calibrating the global latent distribution. By transferring normalized pairwise structure from an informative encoder to the planning latent and applying Wasserstein embedding matching, ATLAS improves goal‑reaching success on tasks such as PushT, TwoRoom, and OGBench‑Cube, especially on higher‑novelty episodes. Diagnostics show stronger novelty‑related structure, better marginal calibration, and lower multi‑step prediction error in the planning latent.
arXiv:2608. 02668v1 Announce Type: cross Abstract: Residual connections are the de facto mechanism for training deep neural networks stably.
Flow-JEPA introduces a conditional flow matching dynamics model that generates a sequence of future latent states conditioned on current observations and actions, replacing deterministic autoregressive prediction with stochastic trajectory-level prediction. By using a Gaussian flow source, the model learns to transport perturbed latent trajectories toward clean future representations while remaining within the reconstruction‑free JEPA framework. The approach improves mean success rates from 86% to 92% under clean observations and from 67% to 86% under noisy conditions.
Geometric foundation models, such as the Visual Geometry Grounded Transformer (VGGT), provide strong 3D priors from unposed images. However, such models operate purely in a feed-forward, deterministic regime, \ie~they cannot generate plausible geometry beyond what the input views directly support.
The paper studies when joint-embedding predictive architectures (JEPAs) can recover underlying causal states from high‑dimensional observations. It introduces a latent variable model where observations arise from causal states with action‑conditioned dynamics, and proposes an information‑theoretic objective that maximizes conditional likelihood while preserving state entropy. The authors prove identifiability conditions—particularly sufficient action‑induced variation—and instantiate the objective as an action‑modulated Gaussian additive‑noise model (A‑JEPA), demonstrating theoretical and empirical success in synthetic and visual benchmarks.
arXiv:2606. 17603v1 Announce Type: new Abstract: In Self-Supervised Learning (SSL), preventing representation collapse by explicitly enforcing a uniform distribution on the unit hypersphere has proven to be effective.
Methods operating on Vision Transformer (ViT) feature spaces typically rely on Euclidean distance or cosine similarity. This assumes that every direction is equally meaningful, but there is no reason...
arXiv:2609. 02155v1 Announce Type: new Abstract: The Johnson-Lindenstrauss (JL) lemma guarantees that a random projection of $n$ points to $m=O(\varepsilon^{-2}\log n)$ dimensions preserves pairwise squared distances within relative error $\varepsilon$ with high probability, and this dimension order is asymptotically optimal.