mlsanityData to 7 Oct 2026

LoGIC: Budgeted Context Construction for Node-Level Graph In-Context Learning with Tabular Foundation Models

arXiv 2609.05955 ↗First posted 5 Sep 2026cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

Tabular foundation models have become powerful graph learners. Systems such as G2T-FM and GraphPFN encode each node as a feature row and make predictions through in-context learning (ICL), with labeled rows serving as the prompt. Current protocols employ the complete training table as context, causing attention to scale quadratically with the labeled pool and introducing preprocessing and memory bottlenecks. We investigate context construction for node-level graph ICL: which labeled nodes and auxiliary unlabeled nodes should constitute the prompt for specified queries. We formulate this allocation in terms of two resources: a labeled-context budget for predictive evidence and an unlabeled-halo budget for adapter message passing without using label capacity. We present LoGIC, which retrieves labeled nodes via structural, feature-based, and coverage channels, shares each context across the queries in a graph-local cluster, incorporates an unlabeled halo for adapter backbones, and chooses the channel and context budget without test labels. Across three backbone configurations drawn from two model families on GraphLand, budgeted contexts maintain locally runnable full-context performance, stay competitive with published large-dataset results, and markedly lower peak memory requirements compared with full-context and whole-graph inference. They further permit frozen graph ICL on million-node graphs without retraining. Our analysis identifies when retrieval channels work best and connects their behavior with graph properties.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 38 of its 43 references that are in the map. First posted on arXiv 5 Sep 2026.

Builds on

38 of 43 references.

References · 38

  1. 2017Inductive Representation Learning on Large Graphscited by 6,284
  2. 2021What Makes Good In-Context Examples for GPT-3?cited by 1,184
  3. 2018Predict then Propagate: Graph Neural Networks meet Personalized PageRankcited by 987
  4. 2022TabPFN: A Transformer That Solves Small Tabular Classification Problems in a Secondcited by 643
  5. 2021Learning To Retrieve Prompts for In-Context Learningcited by 641
  6. 2019Cluster-GCN: An Efficient Algorithm for Training Deep and Large Graph Convolutional Networkscited by 540
  7. 2020Masked Label Prediction: Unified Message Passing Model for Semi-Supervised Classificationcited by 500
  8. 2025TabICL: A Tabular Foundation Model for In-Context Learning on Large Datacited by 267
  9. 2020Combining Label Propagation and Simple Models Out-performs Graph Neural Networkscited by 199
  10. 2020Scaling Graph Neural Networks with Approximate PageRankcited by 174
  11. 2021Graph Condensation for Graph Neural Networkscited by 150
  12. 2023PRODIGY: Enabling In-context Learning Over Graphscited by 97
26 more
  1. 2024TabDPT: Scaling Tabular Foundation Models on Real Datacited by 97
  2. 2022Decoupling the Depth and Scope of Graph Neural Networkscited by 94
  3. 2024GFT: Graph Foundation Model with Transferable Tree Vocabularycited by 82
  4. 2025LimiX: Unleashing Structured-Data Modeling Capability for Generalist Intelligencecited by 65
  5. 2024Retrieval & Fine-Tuning for In-Context Tabular Modelscited by 58
  6. 2017Active Learning for Graph Embeddingcited by 53
  7. 2019Active Learning for Graph Neural Networks via Node Feature Propagationcited by 37
  8. 2024Fully-inductive Node Classification on Arbitrary Graphscited by 36
  9. 2024Mixture of In-Context Prompters for Tabular PFNscited by 32
  10. 2025Turning Tabular Foundation Models into Graph Foundation Modelscited by 30
  11. 2021Grain: Improving Data Efficiency of Graph Neural Networks via Diversified Influence Maximizationcited by 28
  12. 2024RAGraph: A General Retrieval-Augmented Graph Learning Frameworkcited by 24
  13. 2024Let's Ask GNN: Empowering Large Language Model for Graph In-Context Learningcited by 21
  14. 2025GraphICL: Unlocking Graph Learning Potential in LLMs through Structured Prompt Designcited by 21
  15. 2024GraphLand: Evaluating Graph Machine Learning Models on Diverse Industrial Datacited by 20
  16. 2025GraphPFN: A Prior-Data Fitted Graph Foundation Modelcited by 20
  17. 2022Partition-Based Active Learning for Graph Neural Networkscited by 12
  18. 2025Bringing Graphs to the Table: Zero-shot Node Classification via Tabular Foundation Modelscited by 12
  19. 2026Learning Posterior Predictive Distributions for Node Classification from Synthetic Graph Priorscited by 10
  20. 2025Can TabPFN Compete with GNNs for Node Classification via Graph Tabularization?cited by 8
  21. 2026Tabular Foundation Models are Strong Graph Anomaly Detectorscited by 8
  22. 2025Chunked TabPFN: Exact Training-Free In-Context Learning for Long-Context Tabular Datacited by 7
  23. 2025GraphPrompter: Multi-stage Adaptive Prompt Optimization for Graph In-Context Learningcited by 6
  24. 2026TFMLinker: Universal Link Predictor by Graph In-Context Learning with Tabular Foundation Modelscited by 5
  25. 2026A Fair Evaluation of Graph Foundation Models for Node Property Predictioncited by 4
  26. 2026CRUMB: Efficient Prior Fitted Network Inference via Distributionally Matched Context Batchingcited by 2

Built on by

Not cited yet.