English(EN) Beyond Isotropy in JEPAs: Hamiltonian Geometry and Symplectic Prediction

HamJEPA通过哈密顿几何和辛预测推进JEPAs

作者 PulseAugur 编辑部 · [1 个来源] · 2026-05-19 16:57

研究人员推出了一种新颖的联合嵌入预测架构（JEPAs）方法HamJEPA，该方法超越了各向同性正则化。这种新方法将视图编码为相空间状态，并使用学习到的哈密顿跳蛙图进行跨视图预测。在CIFAR-100和ImageNet-100上的实验表明，与SIGReg等现有方法相比，kNN和线性探针精度有了显著提高。 AI

影响引入了一种新的表示学习方法，提高了下游任务的性能。

排序理由该集群包含一篇详细介绍新方法和实验结果的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.AI 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 0 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该集群包含一篇详细介绍新方法和实验结果的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: paper, other; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: 100 days old; Aged out of breaking-news scoring windows; ranking reflects the durable signal from the full source set.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.AI TIER_1 English(EN) · Robert Jenkinson Alvarez · 2026-05-19 16:57

超越JEPAs中的各向同性：哈密顿几何与辛预测

JEPAs often regularize one-view embeddings toward an isotropic Gaussian, implicitly baking Euclidean symmetry into the representation. We show that this is not merely a benign default. For a known structured downstream geometry $H\succ0$, the minimax and maximum-entropy covarianc…

本文如何被排名

报道来源 [1]

超越JEPAs中的各向同性：哈密顿几何与辛预测

相关实体

相关话题