mlsanityData to 7 Oct 2026

InfoNCE is a Free Lunch for Semantically guided Graph Contrastive Learning

arXiv 2505.06282 ↗First posted 7 May 2025cs.LGcited by 3

Not on the map. Why?

Abstract (arXiv)

As an important graph pre-training method, Graph Contrastive Learning (GCL) continues to play a crucial role in the ongoing surge of research on graph foundation models or LLM as enhancer for graphs. Traditional GCL optimizes InfoNCE by using augmentations to define self-supervised tasks, treating augmented pairs as positive samples and others as negative. However, this leads to semantically similar pairs being classified as negative, causing significant sampling bias and limiting performance. In this paper, we argue that GCL is essentially a Positive-Unlabeled (PU) learning problem, where the definition of self-supervised tasks should be semantically guided, i.e., augmented samples with similar semantics are considered positive, while others, with unknown semantics, are treated as unlabeled. From this perspective, the key lies in how to extract semantic information. To achieve this, we propose IFL-GCL, using InfoNCE as a "free lunch" to extract semantic information. Specifically, We first prove that under InfoNCE, the representation similarity of node pairs aligns with the probability that the corresponding contrastive sample is positive. Then we redefine the maximum likelihood objective based on the corrected samples, leading to a new InfoNCE loss function. Extensive experiments on both the graph pretraining framework and LLM as an enhancer show significantly improvements of IFL-GCL in both IID and OOD scenarios, achieving up to a 9.05% improvement, validating the effectiveness of semantically guided. Code for IFL-GCL is publicly available at: https://github.com/Camel-Prince/IFL-GCL.

Key facts. Cited by 3 papers in the map (data through 7 Oct 2026). Builds on 25 of its 47 references that are in the map. First posted on arXiv 7 May 2025.

Builds on

25 of 47 references.

References · 25

  1. 2017Graph Attention Networkscited by 8,854
  2. 2018Representation Learning with Contrastive Predictive Codingcited by 7,662
  3. 2020Understanding Contrastive Representation Learning through Alignment and Uniformity on the Hyperspherecited by 1,393
  4. 2018Deep Graph Infomaxcited by 1,101
  5. 2019DropEdge: Towards Deep Graph Convolutional Networks on Node Classificationcited by 679
  6. 2020Contrastive Multi-View Representation Learning on Graphscited by 651
  7. 2020Graph Contrastive Learning with Adaptive Augmentationcited by 519
  8. 2020GCC: Graph Contrastive Coding for Graph Neural Network Pre-Trainingcited by 461
  9. 2020GPT-GNN: Generative Pre-Training of Graph Neural Networkscited by 302
  10. 2021Graph Contrastive Learning Automatedcited by 287
  11. 2021Self-Supervised Learning of Graph Neural Networks: A Unified Reviewcited by 184
  12. 2022GOOD: A Graph Out-of-Distribution Benchmarkcited by 140
13 more
  1. 2021Self-supervised Heterogeneous Graph Neural Network with Co-contrastive Learningcited by 130
  2. 2021Graph Barlow Twins: A self-supervised representation learning framework for graphscited by 85
  3. 2021HDMI: High-order Deep Multiplex Infomaxcited by 82
  4. 2022Graph Rationalization with Environment-based Augmentationscited by 70
  5. 2024LLMs as Zero-shot Graph Learners: Alignment of GNN Representations with LLM Token Embeddingscited by 60
  6. 2020Self-Supervised Graph Representation Learning via Global Context Predictioncited by 38
  7. 2022COSTA: Covariance-Preserving Feature Augmentation for Graph Contrastive Learningcited by 38
  8. 2024Towards Graph Contrastive Learning: A Survey and Beyondcited by 26
  9. 2020Learning from Positive and Unlabeled Data with Arbitrary Positive Shiftcited by 23
  10. 2024GAugLLM: Improving Graph Contrastive Learning for Text-Attributed Graphs with Large Language Modelscited by 16
  11. 2024Contrastive Learning with Negative Sampling Correctioncited by 4
  12. 2024Negative as Positive: Enhancing Out-of-distribution Generalization for Graph Contrastive Learningcited by 2
  13. 2013Cumulative Effect in Information Diffusion: A Comprehensive Empirical Study on Microblogging Networkcited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 3.

3 papers.

  1. 2026Detecting Miscitation on the Scholarly Web through LLM-Augmented Text-Rich Graph Learningcited by 0
  2. 2026Disentangle-then-Refine: LLM-Guided Decoupling and Structure-Aware Refinement for Graph Contrastive Learningcited by 0
  3. 2026PEARL: Path-Entity Aligned Relational Learning with Contextual Subgraphs for Inductive Knowledge Graph Completioncited by 0