AI Navigate

意味的詳細レベル:双曲多様体上の熱核拡散による多スケール知識表現

arXiv cs.AI / 2026/3/11

Ideas & Deep AnalysisModels & Research

要点

  • 本論文では、双曲多様体上の熱核拡散を用いた知識グラフ表現の連続的解像度制御のためのフレームワークであるSemantic Level of Detail(SLoD)を紹介する。
  • SLoDは粗いスケールで意味的詳細を集約し、細かいスケールで局所的詳細を保持することで、多スケールの知識表現を可能にし、エージェントが抽象化レベルを効果的にナビゲートできるようにする。
  • 本手法は手動パラメータなしでグラフ階層における新たに現れる意味的スケール境界を自動検出し、階層的一貫性と有界近似誤差の理論的証明に支えられている。
  • 実験結果では、SLoDが合成データ上で階層レベルを完全に回復し、WordNet名詞分類体系などの実世界の分類階層とも良く一致することが示された。
  • この手法は、大規模知識グラフにおける意味的抽象化の原理的かつ数学的に根拠のあるメカニズムを提供することで、AIのメモリおよび知識組織の分野を前進させる。

Computer Science > Machine Learning

arXiv:2603.08965 (cs)
[Submitted on 9 Mar 2026]

Title:Semantic Level of Detail: Multi-Scale Knowledge Representation via Heat Kernel Diffusion on Hyperbolic Manifolds

View a PDF of the paper titled Semantic Level of Detail: Multi-Scale Knowledge Representation via Heat Kernel Diffusion on Hyperbolic Manifolds, by Edward Izgorodin
View PDF HTML (experimental)
Abstract:AI memory systems increasingly organize knowledge into graph structures -- knowledge graphs, entity relations, community hierarchies -- yet lack a principled mechanism for continuous resolution control: where do the qualitative boundaries between abstraction levels lie, and how should an agent navigate them? We introduce Semantic Level of Detail (SLoD), a framework that answers both questions by defining a continuous zoom operator via heat kernel diffusion on the Poincaré ball $\mathbb{B}^d$. At coarse scales ($\sigma \to \infty$), diffusion aggregates embeddings into high-level summaries; at fine scales ($\sigma \to 0$), local semantic detail is preserved. We prove hierarchical coherence with bounded approximation error $O(\sigma)$ and $(1+\varepsilon)$ distortion for tree-structured hierarchies under Sarkar embedding. Crucially, we show that spectral gaps in the graph Laplacian induce emergent scale boundaries -- scales where the representation undergoes qualitative transitions -- which can be detected automatically without manual resolution parameters. On synthetic hierarchies (HSBM), our boundary scanner recovers planted levels with ARI up to 1.00, with detection degrading gracefully near the information-theoretic Kesten-Stigum threshold. On the full WordNet noun hierarchy (82K synsets), detected boundaries align with true taxonomic depth ($\tau = 0.79$), demonstrating that the method discovers meaningful abstraction levels in real-world knowledge graphs without supervision.
Comments:
Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
Cite as: arXiv:2603.08965 [cs.LG]
  (or arXiv:2603.08965v1 [cs.LG] for this version)
  https://doi.org/10.48550/arXiv.2603.08965
Focus to learn more
arXiv-issued DOI via DataCite

Submission history

From: Eduard Izgorodin [view email]
[v1] Mon, 9 Mar 2026 21:54:08 UTC (314 KB)
Full-text links:

Access Paper:

    View a PDF of the paper titled Semantic Level of Detail: Multi-Scale Knowledge Representation via Heat Kernel Diffusion on Hyperbolic Manifolds, by Edward Izgorodin
  • View PDF
  • HTML (experimental)
  • TeX Source
Current browse context:
cs.LG
< prev   |   next >
Change to browse by:

References & Citations

export BibTeX citation Loading...

BibTeX formatted citation

×
Data provided by:

Bookmark

BibSonomy logo Reddit logo
Bibliographic Tools

Bibliographic and Citation Tools

Bibliographic Explorer Toggle
Bibliographic Explorer (What is the Explorer?)
Connected Papers Toggle
Connected Papers (What is Connected Papers?)
Litmaps Toggle
Litmaps (What is Litmaps?)
scite.ai Toggle
scite Smart Citations (What are Smart Citations?)
Code, Data, Media

Code, Data and Media Associated with this Article

alphaXiv Toggle
alphaXiv (What is alphaXiv?)
Links to Code Toggle
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub Toggle
DagsHub (What is DagsHub?)
GotitPub Toggle
Gotit.pub (What is GotitPub?)
Huggingface Toggle
Hugging Face (What is Huggingface?)
Links to Code Toggle
Papers with Code (What is Papers with Code?)
ScienceCast Toggle
ScienceCast (What is ScienceCast?)
Demos

Demos

Replicate Toggle
Replicate (What is Replicate?)
Spaces Toggle
Hugging Face Spaces (What is Spaces?)
Spaces Toggle
TXYZ.AI (What is TXYZ.AI?)
Related Papers

Recommenders and Search Tools

Link to Influence Flower
Influence Flower (What are Influence Flowers?)
Core recommender toggle
CORE Recommender (What is CORE?)
IArxiv recommender toggle
IArxiv Recommender (What is IArxiv?)
About arXivLabs

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.