mlsanityData to 14 Oct 2026

Aligning the Spectrum: Hybrid Graph Pre-training and Prompt Tuning across Homophily and Heterophily

arXiv 2508.11328 ↗First posted 15 Aug 2025cs.LG cs.CLcited by 0

Not on the map. Why?

Abstract (arXiv)

Graph ``pre-training and prompt-tuning'' aligns downstream tasks with pre-trained objectives to enable efficient knowledge transfer under limited supervision. However, current methods typically rely on single-filter backbones (e.g., low-pass), whereas real-world graphs exhibit inherent spectral diversity. Our theoretical textitSpectral Specificity principle reveals that effective knowledge transfer requires alignment between pre-trained spectral filters and the intrinsic spectrum of downstream graphs. This identifies two fundamental limitations: (1) Knowledge Bottleneck: single-filter models suffer from irreversible information loss by suppressing signals from other frequency bands (e.g., high-frequency); (2) Utilization Bottleneck: spectral mismatches between pre-trained filters and downstream spectra lead to significant underutilization of pre-trained knowledge. To bridge this gap, we propose HS-GPPT. We utilize a hybrid spectral backbone to construct an abundant knowledge basis. Crucially, we introduce Spectral-Aligned Prompt Tuning to actively align the downstream graph's spectrum with diverse pre-trained filters, facilitating comprehensive knowledge utilization across both homophily and heterophily. Extensive experiments validate the effectiveness under both transductive and inductive learning settings.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 35 of its 45 references that are in the map. First posted on arXiv 15 Aug 2025.

Builds on

35 of 45 references.

References · 35

  1. 2016Semi-Supervised Classification with Graph Convolutional Networkscited by 11,361
  2. 2017Graph Attention Networkscited by 8,923
  3. 2016Convolutional Neural Networks on Graphs with Fast Localized Spectral Filteringcited by 2,089
  4. 2020Graph Contrastive Learning with Augmentationscited by 1,208
  5. 2018Deep Graph Infomaxcited by 1,106
  6. 2020Contrastive Multi-View Representation Learning on Graphscited by 652
  7. 2020Adaptive Universal Generalized PageRank Graph Neural Networkcited by 595
  8. 2019Multi-scale Attributed Node Embeddingcited by 548
  9. 2020Graph Contrastive Learning with Adaptive Augmentationcited by 519
  10. 2020Masked Label Prediction: Unified Message Passing Model for Semi-Supervised Classificationcited by 506
  11. 2022GraphMAE: Self-Supervised Masked Graph Autoencoderscited by 448
  12. 2021Beyond Low-frequency Information in Graph Convolutional Networkscited by 362
23 more
  1. 2023A critical look at the evaluation of GNNs under heterophily: Are we really making progress?cited by 343
  2. 2021BernNet: Learning Arbitrary Graph Spectral Filters via Bernstein Approximationcited by 195
  3. 2022Rethinking Graph Neural Networks for Anomaly Detectioncited by 187
  4. 2024LLaGA: Large Language and Graph Assistantcited by 181
  5. 2022SimGRACE: A Simple Framework for Graph Contrastive Learning without Data Augmentationcited by 171
  6. 2023GraphPrompt: Unifying Pre-Training and Downstream Tasks for Graph Neural Networkscited by 167
  7. 2023All in One: Multi-task Prompting for Graph Neural Networkscited by 148
  8. 2023GraphMAE2: A Decoding-Enhanced Masked Self-Supervised Graph Learnercited by 132
  9. 2020Node Similarity Preserving Graph Convolutional Networkscited by 112
  10. 2020Self-supervised Learning on Graphs: Deep Insights and New Directioncited by 108
  11. 2022Universal Prompt Tuning for Graph Neural Networkscited by 100
  12. 2022Convolutional Neural Networks on Graphs with Chebyshev Approximation, Revisitedcited by 87
  13. 2021AFEC: Active Forgetting of Negative Transfer in Continual Learningcited by 73
  14. 2022Revisiting Graph Contrastive Learning from the Perspective of Graph Spectrumcited by 43
  15. 2023Simple and Asymmetric Graph Contrastive Learning without Augmentationscited by 38
  16. 2023Automated Self-Supervised Learning for Recommendationcited by 37
  17. 2024Non-Homophilic Graph Pre-Training and Prompt Learningcited by 31
  18. 2024ProG: A Graph Prompt Learning Benchmarkcited by 29
  19. 2022EvenNet: Ignoring Odd-Hop Neighbors Improves Robustness of Graph Neural Networkscited by 26
  20. 2025DAGPrompT: Pushing the Limits of Graph Prompting with a Distribution-aware Graph Prompt Tuning Approachcited by 11
  21. 2024Spectral-Based Graph Neural Networks for Complementary Item Recommendationcited by 9
  22. 2022When Does A Spectral Graph Neural Network Fail in Node Classification?cited by 6
  23. 2024Cluster-guided Contrastive Class-imbalanced Graph Classificationcited by 6

Built on by

Not cited yet.