#graphlearning
LightNews LightNews
icelabumu.bsky.social
New research from IceLab members on the #arxiv
#networkscience #graphlearning #scisky
mrosvall.bsky.social Martin Rosvall @mrosvall.bsky.social · Feb 6
We claim that insights from network science can advance deep graph learning in arxiv.org/abs/2502.01177 A fun and enlightening collaboration with Chris Blöcker, Ingo Scholtes, and @jevinwest.bsky.social
Illustrative example of how network science insights can advance deep graph learning.
February 6, 2025 at 12:21 PM
informaq.bsky.social
Implements quantum graph convolutional networks (QSGC, QLGC) with competitive performance using 3log₂C parameters vs. classical C·FK. Proves trainability avoiding barren plateaus for polylog-feature regimes through detailed gradient variance analysis.

#QuantumML #GraphLearning #Research
Quantum Graph Convolutional Networks: Implementation and Trainability Analysis
arxiv.org
September 18, 2026 at 8:07 AM
mpibiochem.bsky.social
We are offering a position as Postdoctoral Research Fellow (M/F/D). Should this position be of interest to you, please apply at the link below:

❕ Apply until: September 30
❕ www.mpg.de/26895302/pos...

#hiring #jobalert #Research #PostDoc #GraphLearning #MachineLearning
August 10, 2026 at 9:41 AM
complexnetconf.bsky.social
⏳ 1 week left to submit to COMPLEX NETWORKS 2026!

📅 Deadline: Sept 2, 2026.
📍 Granada, Spain | Dec 2–4.
📄 Full papers & extended abstracts welcome!
🔗 Submission platform: complexnetworks.org/submission/

#ComplexNetworks #NetworkScience #GraphLearning #ComplexSystems #CallForPapers
August 26, 2026 at 11:22 AM
complexnetconf.bsky.social
⏳ 2 weeks left to submit to #ComplexNetworks2026!

📅 Deadline: Sept 2, 2026
📍 Granada, Spain | Dec 2–4
📄 Full papers & extended abstracts welcome

🔗 complexnetworks.org/submission/
#ComplexNetworks #NetworkScience #GraphLearning #AI #ComplexSystems #Conference #CallForPapers
August 18, 2026 at 11:20 AM
getnews-me.bsky.social
SSTAG combines LLM‑to‑MLP and GNN‑to‑MLP distillation and uses an in‑memory repository of graph anchors for cross‑domain transfer. Read more: https://getnews.me/structure-aware-self-supervised-learning-boosts-text-attributed-graphs/ #graphlearning #selfsupervised #nlp
Structure-Aware Self-Supervised Learning Boosts Text-Attributed Graphs
October 3, 2025 at 7:04 PM
informaq.bsky.social
QDAGer achieves 15% embedding discrepancy reduction using quantum-inspired dynamics to simulate energy propagation, enabling accurate structural analysis of complex graphs previously intractable with classical methods.

#QuantumML #GraphLearning #News
QDAGer: Quantum-Inspired Graph Learning with Dynamic Features
quantumzeitgeist.com
August 22, 2026 at 6:27 PM
informaq.bsky.social
QDAGer leverages quantum dynamics from transverse-field Ising Hamiltonians to generate graph features. Quantum-derived node occupations and correlators outperform classical alternatives for learning graph edit distance on benchmark datasets.

#QuantumML #GraphLearning #Research
Expressive Power of Transverse-Field Ising Models for Graph Learning
arxiv.org
August 19, 2026 at 4:05 AM
lithuania-ai.bsky.social
📣 Call for Papers!
🧠 Neural Networks for Graphs and Beyond @ #ICANN2025
📍 Kaunas, Lithuania | 📅 Deadline: May 1, 2025
Topics: GNNs, temporal graphs, XAI, bio/brain/social data, IoT & more
🔗 Submit: e-nns.org/icann2025/su...
#GNN #GraphLearning #AI #NeuralNetworks #ICANN2025
April 4, 2025 at 1:52 PM
andreaviliotti.bsky.social
𝗖𝗼𝗺𝗲 𝘁𝗿𝗮𝘀𝗳𝗼𝗿𝗺𝗮𝗿𝗲 𝗹𝗮 𝗰𝗼𝗺𝗽𝗹𝗲𝘀𝘀𝗶𝘁𝗮̀ 𝗶𝗻 𝘃𝗮𝗹𝗼𝗿𝗲: 𝗶𝗹 𝗰𝗮𝘀𝗼 𝗚𝗥𝗔𝗙

𝗘 𝘃𝗼𝗶? 𝗖𝗼𝗺𝗲 𝘀𝘁𝗮𝘁𝗲 𝗮𝗳𝗳𝗿𝗼𝗻𝘁𝗮𝗻𝗱𝗼 𝗹𝗮 𝗰𝗼𝗺𝗽𝗹𝗲𝘀𝘀𝗶𝘁𝗮̀ 𝗱𝗲𝗶 𝘃𝗼𝘀𝘁𝗿𝗶 𝗱𝗮𝘁𝗶?

#BigData #GraphLearning #Innovazione #AI #StrategiaDataDriven #DataAnalysis #MachineLearning

www.andreaviliotti.it/post/graf-un...
GRAF: Un nuovo approccio per la fusione di Reti Eterogenee
Il framework GRAF trasforma reti eterogenee e multiplex in omogenee, semplificando l'analisi con tecniche di Graph Representation Learning. Usando meccanismi di attenzione multilivello, GRAF ottimizza...
www.andreaviliotti.it
December 13, 2024 at 9:37 AM
getnews-me.bsky.social
Diffusion-Assisted Distillation for Graph Learning with MLPs (DAD-SGM) uses a diffusion model to bridge GNN-to-MLP knowledge transfer, improving benchmarks. Code on GitHub. Read more: https://getnews.me/diffusion-assisted-distillation-enhances-graph-learning-with-mlps/ #graphlearning #mlp
Diffusion-Assisted Distillation Enhances Graph Learning with MLPs
October 7, 2025 at 11:54 PM
getnews-me.bsky.social
Graphon‑Mixture‑Aware Mixup (GMAM) and Model‑Adaptive GCL (MGCL) boost graph learning; GMAM improves accuracy and MGCL tops average rank on eight benchmark datasets. Read more: https://getnews.me/graphon-mixture-aware-mixup-boosts-contrastive-learning-on-graph-data/ #graphlearning #mixup
Graphon Mixture‑Aware Mixup Boosts Contrastive Learning on Graph Data
October 7, 2025 at 6:12 PM
getnews-me.bsky.social
BrainPoG, a new brain graph learning model, achieved higher accuracy on four benchmark disease detection datasets while using fewer parameters and less processing time. Read more: https://getnews.me/brainpog-lightweight-brain-graph-learning-for-disease-detection/ #brainpog #graphlearning
BrainPoG: Lightweight Brain Graph Learning for Disease Detection
September 29, 2025 at 8:16 AM
getnews-me.bsky.social
CueGCL uses Self‑Training (PeST) and Aligned Graph Clustering (AGC) to cut class collision. Submitted September 2025, it outperforms prior GCL methods on five benchmarks. https://getnews.me/cuegcl-introduces-cluster-aware-self-training-for-unsupervised-graph-learning/ #cuegcl #graphlearning
CueGCL Introduces Cluster-Aware Self-Training for Unsupervised Graph Learning
September 26, 2025 at 11:49 PM
getnews-me.bsky.social
AGCN adds attention to graph edges with a KV cache and pairwise margin contrastive loss, achieving higher silhouette and NMI scores than prior GNN and transformer models. Read more: https://getnews.me/attentive-graph-clustering-network-enhances-transformer-for-graph-data/ #graphlearning #clustering
Attentive Graph Clustering Network Enhances Transformer for Graph Data
September 20, 2025 at 1:50 AM
getnews-me.bsky.social
THGCL (Temporally Heterogeneous Graph Contrastive Learning) improves acoustic event classification, achieving higher precision on the AudioSet benchmark. Paper posted 18 Sep 2025. https://getnews.me/contrastive-graph-learning-improves-multimodal-acoustic-classification/ #graphlearning #audioset
Contrastive Graph Learning Improves Multimodal Acoustic Classification
September 20, 2025 at 12:03 AM
getnews-me.bsky.social
Graph topology learning stays robust with smooth signals from a subset of nodes, recovering edge structure without tweaks. Preprint posted 18 Sep 2025. https://getnews.me/graph-learning-remains-robust-under-partial-observation-of-smooth-signals/ #graphlearning #smoothsignals
Graph Learning Remains Robust Under Partial Observation of Smooth Signals
September 19, 2025 at 11:59 PM
getnews-me.bsky.social
A Bayesian Sheaf Neural Network using a rotation‑group distribution via the Cayley transform achieved leading performance on heterophilic graph benchmarks. Read more: https://getnews.me/bayesian-sheaf-neural-networks-advance-graph-learning-for-heterophilic-data/ #bayesianns #graphlearning
Bayesian Sheaf Neural Networks Advance Graph Learning for Heterophilic Data
September 17, 2025 at 9:22 AM
nithecs.bsky.social
NITheCS & CoRE AI Masterclass: 'An Introduction to Graph Learning & Signal Processing'
🎓 With Dr Fei He & Stephan Goerttler (Coventry University, UK)
🗓️ Tue, 27 May 2025
🕚 11:00–13:00 SAST
📍 Join online or in person
🔗 buff.ly/6nfB5ui

#GraphLearning #SignalProcessing #AI #CoREAI #MachineLearning
May 15, 2025 at 2:21 PM
bifold.berlin
Not just who is connected matters — but when and in which order. ⏳ The arrow of time shapes how nodes influence each other in dynamic networks. This has big implications for graph analytics & deep learning. #AI #GraphLearning
September 22, 2025 at 3:14 PM
arxivlens.bsky.social
Enhancing Homophily-Heterophily Separation: Relation-Aware Learning in
Heterogeneous Graphs
Weigang Lu, Wei Zhao et al.
Paper
Details
#HomophilyHeterophily #GraphLearning #RelationAwareAI
June 30, 2025 at 9:02 AM
iam.slys.dev
Imagine if a neural network could spot not just neighbors but the “style” of a whole graph—cycles, cliques, rare motifs. T-GNNs give GNNs a more nuanced eyesight, changing how machines interpret relational data. How might this shape real-world AI?

GraphLearning AI Motifs Logic
Unifying approach to uniform expressivity of graph neural networks
arxiv.org
June 17, 2026 at 1:54 PM
iam.slys.dev
Unifying graph learning feels overdue. Most real-world data isn’t neatly sorted into simple or complex types. Can modular, expert-driven pre-training help models flex with the messiness of reality? This paper offers one path forward.

unifiedAI graphlearning hybridmodels innovation
Unified Multi-Domain Graph Pre-training for Homogeneous and Heterogeneous Graphs via Domain-Specific Expert Encoding
arxiv.org
April 26, 2026 at 2:19 PM
qbitphased.com
A groundbreaking algorithm blending quantum and classical methods is transforming graph representation learning, unlocking insights into complex networks like social media or biology. How will quantum computing reshape AI's future? 🤔 #QuantumComputing #GraphLearning #AIInnovation LINK
October 7, 2025 at 2:19 PM