Grant_H3d ago

#Higraph progress!

Still got lots to do, but hyperEdges can now be saved & loaded in modified #graphml files. The "model tree" on the left highlights items in the graph on the right.
I can see "minimum viable product"!

The #hyperedge structure is both graphically and algebraically accessible. I'm not aware of anything else that does this, pretty certainly not in #Python
#graphTheory #VisualFormalism

sayzardApr 15

Avi Chawla (@_avichawla)

그래프에서 특정 노드가 영향력 있는 노드들과 연결될수록 더 큰 영향력을 갖게 되는 eigenvector centrality 개념을 설명하고, 코드 예시를 제공한다. 그래프 ML과 네트워크 분석에서 핵심적으로 쓰이는 피처 엔지니어링 기법이다.

https://x.com/_avichawla/status/2044308286363840933

#graphml #eigenvectorcentrality #networkanalysis #featureengineering #machinelearning

Avi Chawla (@_avichawla) on X

6) Eigenvector centrality If a node is connected to other influential nodes, it amplifies its own influence. It helps identify nodes that are influential not only due to their direct ties but also due to their connections with other influential nodes. Here's the code👇

X (formerly Twitter)
sayzardApr 15

Avi Chawla (@_avichawla)

구글 맵, 넷플릭스, 스포티파이, 핀터레스트가 ETA 예측과 추천 시스템에 그래프 ML을 활용하는 사례를 소개하며, 그래프 피처 엔지니어링의 6가지 필수 방법을 코드와 함께 정리한 글이다. 실전 AI/추천 시스템 개발에 유용한 내용이다.

https://x.com/_avichawla/status/2044308125503893792

#graphml #recommendation #featureengineering #machinelearning #ai

Avi Chawla (@_avichawla) on X

- Google Maps uses graph ML to predict ETA - Netflix uses graph ML in recommendation - Spotify uses graph ML in recommendation - Pinterest uses graph ML in recommendation Here are 6 must-know ways for graph feature engineering (with code):

X (formerly Twitter)
Orbifold ConsultingJan 22
Makes my day🙂 Graph ML identifies promising additional drug indications for diseases such as acute lymphoblastic leukemia and Alzheimer’s disease.
https://www.nature.com/articles/s41551-025-01598-z
#KnowledgeGraphs #GraphML #DrugRepurposing
DST GlobalJan 7

LOGOS-κ: Новый язык программирования для моделирования сложных систем

6 января 2026 года Российская компания DST Global и проект Λ-Универсум представили LOGOS-κ — не просто новый язык программирования...

#LOGOS-κ #LOGOSκ #языкпрограммирования #Логос #GraphML #NIGC #DomainSpecificLanguage #DSL #Lambda #Omega #Универсум #Universum #ΛУниверсум #AUniversum #АУниверсум #Искусственныйинтеллект

Ссылки:
Репозиторий: https://github.com/A-Universum/logos-k
Исходный манифест Λ-Универсума: https://github.com/a-universum

Orbifold ConsultingNov 13
Node and graph embeddings are a bit of a burden if you want to do simple experiments: PyG (aka Torch Geometric) comes with a steep learning curve. This NEExT package is a lot easier and supports NetworkX and iGraph, Wasserstein and more.
https://github.com/ashdehghan/NEExT
https://skolouri.github.io/wegl/
#GraphML #GraphAI
HackerNoonSep 11, 2025
This article tests how degree, clustering, and topology–feature ties sway GNN and feature-only models using HypNF synthetic graphs. https://hackernoon.com/choose-the-right-graph-model-faster-with-hypnfs-parameter-knobs #graphml
Choose the Right Graph Model Faster with HypNF’s Parameter Knobs | HackerNoon

This article tests how degree, clustering, and topology–feature ties sway GNN and feature-only models using HypNF synthetic graphs.

HackerNoonSep 10, 2025
Synthetic HypNF graphs reveal GNN fragilities: HGCN beats GCN on dense, homogeneous nets but falters on sparse power-law ones. https://hackernoon.com/a-hyperbolic-benchmark-for-stress-testing-gnns-across-degree-clustering-and-homophily #graphml
A Hyperbolic Benchmark for Stress-Testing GNNs Across Degree, Clustering, and Homophily | HackerNoon

Synthetic HypNF graphs reveal GNN fragilities: HGCN beats GCN on dense, homogeneous nets but falters on sparse power-law ones.

SadamoriMay 28, 2024
🚨Preprint Alert🚨 Benchmarks guide #MachineLearning, but is the core benchmark for #GraphML, the link prediction task, guiding us correctly? Our latest preprint questions its validity, reveals a bias that substantially skews the evaluation, and proposes a degree-corrected link prediction benchmark. Let's dive in! https://arxiv.org/abs/2405.14985 More on 👉 https://twitter.com/skojaku/status/1795413358818013431
Implicit degree bias in the link prediction task

Link prediction -- a task of distinguishing actual hidden edges from random unconnected node pairs -- is one of the quintessential tasks in graph machine learning. Despite being widely accepted as a universal benchmark and a downstream task for representation learning, the validity of the link prediction benchmark itself has been rarely questioned. Here, we show that the common edge sampling procedure in the link prediction task has an implicit bias toward high-degree nodes and produces a highly skewed evaluation that favors methods overly dependent on node degree, to the extent that a ``null'' link prediction method based solely on node degree can yield nearly optimal performance. We propose a degree-corrected link prediction task that offers a more reasonable assessment that aligns better with the performance in the recommendation task. Finally, we demonstrate that the degree-corrected benchmark can more effectively train graph machine-learning models by reducing overfitting to node degrees and facilitating the learning of relevant structures in graphs.

arXiv.org
meOct 30, 2023
@tml @Sweetshark @boris

used yEd couple times. Very flexible and export functions.
Liked their layout algorithms for objects en connector routing styles.

https://www.yworks.com/products/yed

#yed #graph #graphml
yEd Graph Editor

yEd is a free desktop application to quickly create, import, edit, and automatically arrange diagrams. It runs on Windows, macOS, and Unix/Linux.

yWorks, the diagramming experts