Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

Pre-training Graph Neural Networks on Molecules by Using Subgraph-Conditioned Graph Information Bottleneck

T0 review · 3 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Pre-training on unlabeled molecules by compressing each molecule into a core and reconstructing it from attention-selected local subgraphs yields representations that, the paper reports, outperform previous methods on 10 of 11…

desk verdict S-CGIB is a solid, honest molecular pre-training paper whose headline claim overstates the contribution once you subtract the domain-adaptation stage that baselines don't get. read the letter →

arxiv 2412.15589 v1 pith:6ZX2ZOL7 submitted 2024-12-20 cs.LG cs.AI

Van Thuy Hoang , O-Joun Lee This is my paper · ORCID
classification cs.LGcs.AI
keywords graphneuralnetworksself-supervisedpre-trainingmolecularpropertypredictioninformationbottleneckfunctionalgroupdiscoveryego-networksubgraphsgraph-levelrepresentationlearning
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Pre-training a graph neural network on molecules usually leans on hand-annotated functional groups or mined fragment vocabularies. This paper proposes a self-supervised alternative in which the network compresses each molecule into a graph core and then must reconstruct the molecule from that core together with attention-selected ego-network subgraphs. The reported result is that this joint compression-and-reconstruction objective produces graph-level representations that transfer well: it wins 10 of 11 classification benchmarks and matches or beats prior methods on regression and large-molecule tasks. The paper further reports that the attention weights single out subgraphs that agree with ground-truth functional groups, even though no functional-group labels were used during pre-training. If true, this would remove the need for curated fragment dictionaries and make pre-training cover rare or novel functional groups automatically.

What carries the argument

The machine that carries the argument is the S-CGIB objective, a conditional variant of the graph information bottleneck in which the side information is the set $S=\{G[N_k(v)] \mid v\in V\}$ of $k$-hop ego-network functional-group candidates. The two terms are made trainable in the following way: reconstruction is implemented as a cosine-similarity adjacency-matrix recovery loss from representations that concatenate the compressed core with attention-weighted ego-network embeddings, and compression is implemented by sampling Bernoulli masks over node embeddings with a variational upper bound plus a contrastive alignment term. The attention-based interaction between the pooled core and each ego-network embedding is what selects significant subgraphs, and it is this selection mechanism, not any external fragment dictionary, that is claimed to discover functional groups.

What would settle it

A randomized-subgraph control: replace the attention-selected ego networks with equally sized subgraphs drawn at random, retrain, and compare downstream accuracy and fidelity scores; if performance is unchanged, the specific local structure of ego networks is not what carries the gain and the functional-group interpretation loses support.

Watch

Extended reading notes

Core claim

On its own terms, the discovery is that the label-dependent graph information bottleneck can be converted into a self-supervised objective by conditioning compression on local subgraphs. The S-CGIB objective is $$\min_{G_c} -I(G; G_c|S) + \$\beta$ I(G; G_c),$$ where $S$ is the set of $k$-hop ego networks rooted at each atom and $G_c$ is a core obtained by Bernoulli noise-masking of node embeddings. The first term becomes adjacency-matrix reconstruction from core-plus-attended-subgraph representations, and the second term is minimized through a variational upper bound, with a contrastive term to prevent over-compression. Across eleven graph-classification datasets the paper reports the best result in ten, and it reports gains on regression and two large peptide benchmarks, while the learned attention coefficients mark subgraphs that align with known functional groups in interpretability datasets. This is the paper's evidence that both well-separated graph representations and functional groups can be obtained from unlabeled molecular graphs alone.

Load-bearing premise

The load-bearing premise is that a molecule is faithfully decomposable into a small core plus the collection of its $k$-hop ego networks, because the reconstruction loss is what teaches the model which subgraphs act as functional groups.

Editorial extensions

If this is right

  • Molecular pre-training can drop manually curated functional-group dictionaries and frequent-subgraph mining, because the attention mechanism builds the group vocabulary from unlabeled molecules.
  • One pre-trained S-CGIB checkpoint on 300k unlabeled molecules can be fine-tuned across biophysics, physiology, physical chemistry, bioinformatics, and large-peptide benchmarks, lowering the cost of adapting to new tasks.
  • The attention weights themselves are explanations: on datasets with ground-truth labels, the top-ranked ego networks match known functional groups, so downstream predictions become inspectable.
  • Compression plus contrastive alignment produces more separated class-level embeddings than prior methods, which should make downstream classifiers easier to train from few labels.
  • The reported improvement on large peptide graphs over fragment-dictionary methods indicates the core-plus-subgraph objective captures structure beyond small frequent fragments.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Replacing the ego-network candidate set with equally sized random subgraphs would be a decisive control: if downstream performance and fidelity scores barely move, the specific locality of ego networks is not what drives the reported gains.
  • Since the best results appear at $k \le 3$, letting the radius adapt per node or mixing in ring-based fragments might capture functional groups that are larger or non-local, which the current fixed-radius candidate set cannot represent.
  • The Table 7 ablation suggests the attention-selected subgraph channel is at least as valuable as the compressed core, so a simpler model that keeps the subgraph attention but drops noise-injection compression might match S-CGIB and shift the explanation from information-theoretic compression to subgraph saliency.
  • Dropping $I(G;S)$ from Eq. 14 is justified empirically in the paper; a version that keeps a small weight on that term, or that estimates it more carefully, could yield a more stable objective and better out-of-domain transfer.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper proposes S-CGIB, a self-supervised pre-training method for molecular graph neural networks. The method compresses an input molecule into a graph core by injecting noise into node embeddings, extracts k-hop ego networks rooted at each node as functional-group candidates, and uses an attention mechanism between the graph core and these candidates to reconstruct the input adjacency matrix. The pre-training loss combines a conditional-reconstruction term, a variational upper bound on the compression mutual information I(G;Gc), and a contrastive term. After pre-training, an unsupervised domain-adaptation stage reconstructs node features on downstream data. The method is evaluated on 11 graph classification, 3 regression, and 2 long-range graph benchmarks, along with interpretability datasets, and is claimed to outperform prior pre-training baselines consistently.

Significance. The empirical effort is substantial and, in large part, honestly reported: the paper covers a wide range of benchmarks, includes ablations and sensitivity analyses, reports results both with and without the domain-adaptation stage, and makes code available. If the theoretical framing were faithful and the comparisons controlled, the automatic discovery of functional groups through attention-based interaction with a compressed core would be a useful contribution to molecular pre-training. However, the objective actually minimized is not the S-CGIB objective stated in Definition 3, and the headline performance claim is entangled with an extra domain-adaptation stage that baselines do not receive. The significance of the claimed gains is therefore not yet established in the current form.

major comments (3)
  1. [Model Optimization, Eq. (14); Appendix D.1] The chain-rule identity in Eq. (14) is correct, but the optimization that follows does not implement it: dropping I(G;S) changes the objective from -I(G;Gc|S)+βI(G;Gc) to -I(G;Gc,S)+βI(G;Gc). The empirical degradation reported in Appendix D.1 motivates this choice, but it is not a derivation, and it means the central claim of optimizing a 'Subgraph-conditioned Graph Information Bottleneck' is not supported by the actual objective. Please either redefine S-CGIB to be the objective that is truly optimized, or provide a principled argument for why I(G;S) can be neglected.
  2. [Model Optimization, Eqs. (17)-(19)] The actual pre-training loss is Ltotal = Lcon + Lrec + βLMI, but the contrastive term Lcon does not appear in Eq. (13) or in Definition 3. No weight or derivation is given for Lcon, and maximizing agreement between the graph core and the full input graph is in tension with the compression term βI(G;Gc). Since Lcon is part of the method, the statement that S-CGIB optimizes the information-bottleneck objective in Eq. (13) is incomplete; please incorporate Lcon into the derivation or describe it explicitly as a separate heuristic component.
  3. [Evaluation, Tables 1-2; Eq. (20)] The headline claim that S-CGIB 'consistently outperformed other baselines, obtaining the best performance in 10 out of 11 downstream datasets' refers to the full pipeline that includes the unsupervised domain-adaptation stage Latt, which is not applied to any baseline. The S-CGIB w/o D.A. rows do not show consistent superiority: on BBBP the score is 86.71 versus 87.15 for GROVE; on SIDER it is 62.76 versus 65.93 for GraphFP; on ClinTox it is 74.69 versus 77.56 for MICRO-Graph; on HIV it is 77.41 versus 78.13 for SimSGT; and on Mutagenicity it is 80.26 versus 80.64 for MICRO-Graph. Because Latt is an uncontrolled additional stage, the incremental value of the S-CGIB pre-training loss itself is not isolated. Please add baselines with an analogous domain-adaptation stage, or restrict the superiority claim to the w/o D.A. comparison.
minor comments (5)
  1. [Interpretability Analysis, Table 4] Fidelity is a model-behavior faithfulness metric; by itself it does not show agreement with ground-truth functional groups. The qualitative examples in Figure 3 are the direct evidence for the 'match the real-world functional groups' claim, and the text should state this distinction.
  2. [Related Work and Tables 1-2] The method name GROVER in the related-work section is written as GROVE in Tables 1 and 2; please make the notation consistent.
  3. [Appendix B] There are several typos: 'Carbony' should be 'Carbonyl', 'benzene zing' should be 'benzene ring', and 'F idelity' in Table 4 should be 'Fidelity'.
  4. [Problem Descriptions] The statement G = Gc ∪ S is not formally defined. Since S is already the set of all k-hop ego networks, the union operation over graphs needs a precise definition; otherwise the assumption is close to tautological.
  5. [Appendix D.1, Eq. (29)] In Eq. (29), the mutual information I(G;S) is replaced by a batch-averaged dot product; the paper should state explicitly that this is a heuristic approximation of the mutual information used in Eq. (14), not the information-theoretic quantity itself.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: the S-CGIB objective is self-supervised, its benchmarks are external, and the few self-citations are not load-bearing.

full rationale

The paper's pre-training objective (Eq. 13) is not constructed from downstream labels or from the benchmark outcomes; it minimizes a reconstruction loss, a compression bound, and a contrastive term on unlabeled molecules, and the downstream evaluations are external datasets. The functional-group discovery is an emergent product of attention weights trained by graph reconstruction, and it is validated post hoc against external ground-truth explanations (Table 4 and Figure 3), so no fitted parameter is renamed as a prediction. The GIB and VGIB upper bounds are imported from external prior work (Yu et al. 2021; Yu, Cao, and He 2022; Alemi et al. 2017), not from the authors' own prior papers. The only self-citations (Hoang et al. 2023; Hoang and Lee 2024) occur in related-work context and do not carry the derivation. One acknowledged derivation gap is the dropping of I(G;S) in Eq. 14, which is justified empirically by the ablation in Appendix D.1 rather than by a bound; this weakens the formal derivation but is not circular. A separate experimental concern, namely that the domain-adaptation stage (Eq. 20) is applied only to S-CGIB, affects the comparability of the headline 10-of-11 claim, but that is a protocol issue rather than a circularity of the derivation. Overall, no load-bearing step reduces to its own input, so no significant circularity is found.

Assumptions & free parameters 5 free parameters · 5 assumptions · 2 invented entities

The central claim rests on the decomposition G = Gc ∪ S and on several variational approximations. The dropped I(G;S) term is the most consequential ad hoc choice. No new physical entities are introduced; 'graph core' and 'significant subgraphs' are model constructs.

free parameters (5)
  • beta (compression weight) = 1.0 (grid-searched over 0.01, 0.1, 1.0, 10)
    Balances L_MI against reconstruction and contrastive losses; tuned on downstream tasks.
  • tau (Gumbel sigmoid temperature) = 1.0 (tuned over 1.0, 0.5, 0.1)
    Sharpness of the Bernoulli relaxation for core-node selection.
  • k (ego-network hop size) = not fixed in main tables; sensitivity shows k <= 3 best
    Defines the functional-group candidate subgraphs; reported results depend on this choice.
  • embedding dimension = 64 (searched over 32, 64, 128, 256)
    GNN representation size; reported results use 64.
  • number of GIN layers = 5
    Encoder depth; sensitivity analysis shows performance stable at 3 or more layers.
assumptions (5)
  • standard math Non-negativity of KL divergence used to justify variational upper bounds (Eq. 15 and Eq. 25).
    Appendix A.1 and A.2 rely on KL(p,q) >= 0 to derive the bounds.
  • domain assumption Molecules decompose as G = Gc ∪ S, with Gc a universal core and S the set of k-hop ego networks.
    Stated in Problem Descriptions: 'we suppose that G is formed by combining a graph core Gc and a set of functional group candidates S'. Reconstruction from a small core plus ego networks is assumed sufficient to recover the molecule.
  • ad hoc to paper The term I(G;S) in the chain-rule decomposition of -I(G;Gc|S) is dropped from the optimization.
    Justified only empirically in Appendix D.1 (adding it hurts performance); it changes the objective from -I(G;Gc|S) to -I(G;Gc,S), a deviation specific to this paper.
  • domain assumption The variational decoder p_zeta(G|Gc,S) is modeled as cosine-similarity adjacency reconstruction (Eq. 16).
    Assumes Frobenius loss on the adjacency matrix is a sufficient surrogate for graph likelihood.
  • domain assumption Sufficient encoder assumption I(Z|H) ~ I(Gc|G) holds.
    Invoked in Appendix A.2 to connect the MI upper bound to the encoder output; unproved, standard in VGIB.
invented entities (2)
  • Graph core Gc
    purpose: Compressed latent subgraph that preserves sufficient information for reconstruction and generates well-separated graph representations.
    Implemented by noise injection and Bernoulli masking; no falsifiable handle outside the model.
  • Significant subgraphs (functional-group candidates)
    purpose: Side information for reconstructing the input graph and for automatically discovering functional groups via attention.
    Selected from k-hop ego networks; matches to chemical functional groups are evaluated post hoc, so the construct itself is not independently observable.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Pre-training Graph Neural Networks on Molecules by Using Subgraph-Conditioned Graph Information Bottleneck." pith.science (2026). https://pith.science/paper/6ZX2ZOL7

@misc{pith2026241215589,
  author       = {Pith},
  title        = {Pith review of: Pre-training Graph Neural Networks on Molecules by Using Subgraph-Conditioned Graph Information Bottleneck},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6ZX2ZOL7}},
  note         = {Machine review of arXiv:2412.15589}
}
read the original abstract

This study aims to build a pre-trained Graph Neural Network (GNN) model on molecules without human annotations or prior knowledge. Although various attempts have been proposed to overcome limitations in acquiring labeled molecules, the previous pre-training methods still rely on semantic subgraphs, i.e., functional groups. Only focusing on the functional groups could overlook the graph-level distinctions. The key challenge to build a pre-trained GNN on molecules is how to (1) generate well-distinguished graph-level representations and (2) automatically discover the functional groups without prior knowledge. To solve it, we propose a novel Subgraph-conditioned Graph Information Bottleneck, named S-CGIB, for pre-training GNNs to recognize core subgraphs (graph cores) and significant subgraphs. The main idea is that the graph cores contain compressed and sufficient information that could generate well-distinguished graph-level representations and reconstruct the input graph conditioned on significant subgraphs across molecules under the S-CGIB principle. To discover significant subgraphs without prior knowledge about functional groups, we propose generating a set of functional group candidates, i.e., ego networks, and using an attention-based interaction between the graph core and the candidates. Despite being identified from self-supervised learning, our learned subgraphs match the real-world functional groups. Extensive experiments on molecule datasets across various domains demonstrate the superiority of S-CGIB.

Figures

Figures reproduced from arXiv: 2412.15589 by the authors.

Figure 1
Figure 1. The overall architecture of S-CGIB. where g (l) θ (·) denotes a GNN encoder, e.g., GIN, for the sub￾graph G[Nk(v)], POOL(·) refers to a subgraph pooling, e.g., sum. Therefore, the local surrounding structures rooted at each node are captured via GNNs. Graph Core and Subgraph Interaction Note that each functional group has distinct chemical characteristics that contribute differently to the overall molecule behavior.… view at source ↗
Figure 2
Figure 2. An efficiency analysis for variants of S-CGIB. The [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Visualizations of model interpretability in functional group detection tasks. [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Performance according to weighting factor [PITH_FULL_IMAGE:figures/full_fig_p014_4.png]
Figure 5
Figure 5. Figure 5: Performance according to subgraph sizes ( [PITH_FULL_IMAGE:figures/full_fig_p015_5.png]
Figure 6
Figure 6. Figure 6: Performance according to the number of GIN lay [PITH_FULL_IMAGE:figures/full_fig_p015_6.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Mitigating Degree Bias in Graph Representation Learning with Learnable Structural Augmentation and Structural Self-Attention

    cs.AI 2025-04 conditional novelty 5.0 of 10

    DegFairGT adds degree-weighted edges between similar non-adjacent nodes within communities and injects structural similarity into self-attention, improving degree fairness, classification, and clustering on six benchmarks.

Reference graph

Works this paper leans on

53 extracted references · 45 canonical work pages · cited by 1 Pith paper

  1. [1]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

  3. [3]

    Agarwal, C.; Queen, O.; Lakkaraju, H.; and Zitnik, M. 2023. Evaluating explainability for graph neural networks. Scientific Data, 10(1): 144

    work page 2023

  4. [4]

    A.; Fischer, I.; Dillon, J

    Alemi, A. A.; Fischer, I.; Dillon, J. V.; and Murphy, K. 2017. Deep Variational Information Bottleneck. In 5th International Conference on Learning Representations ( ICLR 2017) . Toulon, France: OpenReview.net

    work page 2017

  5. [5]

    Amara, K.; Ying, Z.; Zhang, Z.; Han, Z.; Zhao, Y.; Shan, Y.; Brandes, U.; Schemm, S.; and Zhang, C. 2022. GraphFramEx: Towards Systematic Evaluation of Explainability Methods for Graph Neural Networks. In Proceedings of the 1st Learning on Graphs Conference (LoG 2022). Virtual Event

    work page 2022

  6. [6]

    Cao, S.; Lu, W.; and Xu, Q. 2015. GraRep: Learning Graph Representations with Global Structural Information. In 24th International Conference on Information and Knowledge Management ( CIKM 2015) , 891--900. Melbourne, VIC, Australia: ACM

    work page 2015

  7. [7]

    Chechik, G.; and Tishby, N. 2002. Extracting Relevant Structures with Side Information. In Proceedings of the 15th Advances in Neural Information Information Processing Systems (NIPS 2002), 857--864. Vancouver, British Columbia, Canada: MIT Press

    work page 2002

  8. [8]

    Degen, J.; Wegscheid-Gerlach, C.; Zaliani, A.; and Rarey, M. 2008. On the art of compiling and using'drug-like'chemical fragment spaces. ChemMedChem, 3(10): 1503

    2008

Show all 53 references
  1. [9]

    Devlin, J.; Chang, M.; Lee, K.; and Toutanova, K. 2019. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. In Proceedings of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies ( NAACL-HLT 2019...

    2019

  2. [10]

    P.; and Bresson, X

    Dwivedi, V. P.; and Bresson, X. 2021. A Generalization of Transformer Networks to Graphs. In Proceedings of the AAAI Workshop on Deep Learning on Graphs: Methods and Applications (AAAIW 2021)

    2021

  3. [11]

    P.; Ramp \' a sek, L.; Galkin, M.; Parviz, A.; Wolf, G.; Luu, A

    Dwivedi, V. P.; Ramp \' a sek, L.; Galkin, M.; Parviz, A.; Wolf, G.; Luu, A. T.; and Beaini, D. 2022. Long Range Graph Benchmark. In Proceedings of the 35th Advances in Neural Information Processing System (NeurIPS 2022). New Orleans, LA, USA

    2022

  4. [12]

    Fey, M.; and Lenssen, J. E. 2019. Fast Graph Representation Learning with PyTorch Geometric . In Proceedings of the ICLR Workshop on Representation Learning on Graphs and Manifolds (ICLRW 2019)

    2019

  5. [13]

    S.; Riley, P

    Gilmer, J.; Schoenholz, S. S.; Riley, P. F.; Vinyals, O.; and Dahl, G. E. 2017. Neural Message Passing for Quantum Chemistry. In Proceedings of the 34th International Conference on Machine Learning ( ICML 2017) , volume 70 of PMLR, 1263--1272. Sydney, NSW, Australia: PMLR

    2017

  6. [14]

    Gondek, D.; and Hofmann, T. 2003. Conditional information bottleneck clustering. In Proceedings of the 3rd IEEE international conference on data mining, workshop on clustering large data sets (ICMDW 2003), 36--42

    2003

  7. [15]

    L.; Ying, Z.; and Leskovec, J

    Hamilton, W. L.; Ying, Z.; and Leskovec, J. 2017 a . Inductive Representation Learning on Large Graphs. In Proceedings of the 30th Annual Conference on Neural Information Processing Systems (NeurIPS 2017), 1024--1034. Long Beach, CA, USA

    2017

  8. [16]

    L.; Ying, Z.; and Leskovec, J

    Hamilton, W. L.; Ying, Z.; and Leskovec, J. 2017 b . Inductive Representation Learning on Large Graphs. In Proceedings of the 30th Annual Conference on Neural Information Processing Systems (NeurIPS 2017), 1024--1034. Long Beach, CA, USA

    2017

  9. [17]

    Hao, Z.; Lu, C.; Huang, Z.; Wang, H.; Hu, Z.; Liu, Q.; Chen, E.; and Lee, C. 2020. ASGN: An Active Semi-supervised Graph Neural Network for Molecular Property Prediction. In Proceedings of the 26th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD 2020) , 731--...

    2020

  10. [18]

    T.; Jeon, H.-J.; You, E.-S.; Yoon, Y.; Jung, S.; and Lee, O.-J

    Hoang, V. T.; Jeon, H.-J.; You, E.-S.; Yoon, Y.; Jung, S.; and Lee, O.-J. 2023. Graph Representation Learning and Its Applications: A Survey. Sensors, 23(8)

    2023

  11. [19]

    T.; and Lee, O

    Hoang, V. T.; and Lee, O. 2024. Transitivity-Preserving Graph Representation Learning for Bridging Local Connectivity and Role-Based Similarity. In Proceedings of the 38th Conference on Artificial Intelligence ( AAAI 2024) , 12456--12465. Vancouver,Canada: AAAI Press

    2024

  12. [20]

    Hu, W.; Fey, M.; Ren, H.; Nakata, M.; Dong, Y.; and Leskovec, J. 2021. OGB-LSC: A Large-Scale Challenge for Machine Learning on Graphs. In Proceedings of the 1st Neural Information Processing Systems Track on Datasets and Benchmarks (NeurIPS 2021). Virtual Event

    2021

  13. [21]

    S.; and Leskovec, J

    Hu, W.; Liu, B.; Gomes, J.; Zitnik, M.; Liang, P.; Pande, V. S.; and Leskovec, J. 2020 a . Strategies for Pre-training Graph Neural Networks. In Proceedings of the 8th International Conference on Learning Representations ( ICLR 2020) . Addis Ababa, Ethiopia: OpenReview.net

    2020

  14. [22]

    Hu, Z.; Dong, Y.; Wang, K.; Chang, K.; and Sun, Y. 2020 b . GPT-GNN: Generative Pre-Training of Graph Neural Networks. In Proceedings of the 26th SIGKDD Conference on Knowledge Discovery and Data Mining (KDD 2020) , 1857--1867. Virtual Event: ACM

    2020

  15. [23]

    Inae, E.; Liu, G.; and Jiang, M. 2023. Motif-aware Attribute Masking for Molecular Graph Pre-training. arXiv preprint, arXiv:2309.04589

    2023 arXiv

  16. [24]

    Jang, E.; Gu, S.; and Poole, B. 2017. Categorical Reparameterization with Gumbel-Softmax. In Proceedings of the 5th International Conference on Learning Representations ( ICLR 2017) . Toulon, France: OpenReview.net

    2017

  17. [25]

    Kazius, J.; McGuire, R.; and Bursi, R. 2005. Derivation and validation of toxicophores for mutagenicity prediction. Journal of medicinal chemistry, 48(1): 312--320

    2005

  18. [26]

    S.; and Riley, P

    Kearnes, S.; McCloskey, K.; Berndl, M.; Pande, V. S.; and Riley, P. 2016. Molecular graph convolutions: moving beyond fingerprints. Journal of Computer-Aided Molecular Design, 30(8): 595--608

    2016

  19. [27]

    N.; and Welling, M

    Kipf, T. N.; and Welling, M. 2017. Semi-Supervised Classification with Graph Convolutional Networks. In Proceedings of the 5th International Conference on Learning Representations ( ICLR 2017) . Toulon, France: OpenReview.net

    2017

  20. [28]

    Kong, X.; Huang, W.; Tan, Z.; and Liu, Y. 2022. Molecule Generation by Principal Subgraph Mining and Assembling. In Proceedings of the 35th Advances in Neural Information Processing Systems (NeurIPS 2022). New Orleans, LA, USA

    2022

  21. [29]

    Lee, N.; Lee, J.; and Park, C. 2022. Augmentation-Free Self-Supervised Learning on Graphs. In Proceedings of the 36th Conference on Artificial Intelligence ( AAAI 2022) , 7372--7380. Virtual Event: AAAI Press

    2022

  22. [30]

    Liu, S.; Wang, H.; Liu, W.; Lasenby, J.; Guo, H.; and Tang, J. 2022. Pre-training Molecular Graph Representation with 3D Geometry. In Proceedings of the 10th International Conference on Learning Representations ( ICLR 2022) . Virtual Event: OpenReview.net

    2022

  23. [31]

    Liu, Z.; Shi, Y.; Zhang, A.; Zhang, E.; Kawaguchi, K.; Wang, X.; and Chua, T. 2023. Rethinking Tokenizer and Decoder in Masked Graph Modeling for Molecules. In Proceedings of the 36th Advances in Neural Information Processing Systems (NeurIPS 2023). New Orleans, LA, USA

    2023

  24. [32]

    Luong, K.; and Singh, A. K. 2023. Fragment-based Pretraining and Finetuning on Molecular Graphs. In Proceedings of the 36th Advances in Neural Information Processing Systems (NeurIPS 2023). New Orleans, LA, USA

    2023

  25. [33]

    J.; Mnih, A.; and Teh, Y

    Maddison, C. J.; Mnih, A.; and Teh, Y. W. 2017. The Concrete Distribution: A Continuous Relaxation of Discrete Random Variables. In Proceedings of the 5th International Conference on Learning Representations (ICLR 2017). Toulon, France

    2017

  26. [34]

    M.; Bause, F.; Kersting, K.; Mutzel, P.; and Neumann, M

    Morris, C.; Kriege, N. M.; Bause, F.; Kersting, K.; Mutzel, P.; and Neumann, M. 2020. TUDataset: A collection of benchmark datasets for learning with graphs. In Proceedings of the ICML 2020 Workshop on Graph Representation Learning and Beyond (ICMLW 2020)

    2020

  27. [35]

    E.; Kolouri, S.; Rostami, M.; Martin, C

    Pope, P. E.; Kolouri, S.; Rostami, M.; Martin, C. E.; and Hoffmann, H. 2019. Explainability Methods for Graph Convolutional Neural Networks. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition ( CVPR 2019) , 10772--10781. Long Beach, CA, USA: Compu...

    2019

  28. [36]

    Qiu, J.; Chen, Q.; Dong, Y.; Zhang, J.; Yang, H.; Ding, M.; Wang, K.; and Tang, J. 2020. GCC: Graph Contrastive Coding for Graph Neural Network Pre-Training. In Proceedings of the Conference on Knowledge Discovery and Data Mining (KDD 2020), 1150--1160. Virtual Event: ACM

    2020

  29. [37]

    Rong, Y.; Bian, Y.; Xu, T.; Xie, W.; Wei, Y.; Huang, W.; and Huang, J. 2020. Self-Supervised Graph Transformer on Large-Scale Molecular Data. In Proceedings of the 34th Annual Conference on Neural Information Processing Systems (NeurIPS 2020). Virtual Event

    2020

  30. [38]

    a rk, H.; Beaini, D.; Corso, G.; Tossou, P.; Dallago, C.; G \

    St \" a rk, H.; Beaini, D.; Corso, G.; Tossou, P.; Dallago, C.; G \" u nnemann, S.; and Li \' o , P. 2022. 3D Infomax improves GNNs for Molecular Property Prediction. In Proceedings of the International Conference on Machine Learning ( ICML 2022) , volume 162 of PMLR, 20479--2...

    2022

  31. [39]

    Sterling, T.; and Irwin, J. J. 2015. ZINC 15--ligand discovery for everyone. Journal of chemical information and modeling, 55(11): 2324--2337

    2015

  32. [40]

    Subramonian, A. 2021. MOTIF-Driven Contrastive Learning of Graph Representations. In Proceedings of the 35th Conference on Artificial Intelligence ( AAAI 2021) , 15980--15981. Virtual Event: AAAI Press

    2021

  33. [41]

    Tishby, N.; Pereira, F. C. N.; and Bialek, W. 2000. The information bottleneck method. arXiv preprint, arXiv:physics-0004057

    2000

  34. [42]

    Velickovic, P.; Cucurull, G.; Casanova, A.; Romero, A.; Li \` o , P.; and Bengio, Y. 2018. Graph Attention Networks. In Proceedings of the 6th International Conference on Learning Representations ( ICLR 2018) . Vancouver, BC, Canada: OpenReview.net

    2018

  35. [43]

    L.; Li \` o , P.; Bengio, Y.; and Hjelm, R

    Velickovic, P.; Fedus, W.; Hamilton, W. L.; Li \` o , P.; Bengio, Y.; and Hjelm, R. D. 2019. Deep Graph Infomax. In Proceedings of the 7th International Conference on Learning Representations ( ICLR 2019) . New Orleans, LA, USA: OpenReview.net

    2019

  36. [44]

    Wang, M.; Zheng, D.; Ye, Z.; Gan, Q.; Li, M.; Song, X.; Zhou, J.; Ma, C.; Yu, L.; Gai, Y.; Xiao, T.; He, T.; Karypis, G.; Li, J.; and Zhang, Z. 2019. Deep Graph Library: A Graph-Centric, Highly-Performant Package for Graph Neural Networks. arXiv preprint arXiv:1909.01315

    2019 arXiv

  37. [45]

    N.; Gomes, J.; Geniesse, C.; Pappu, A

    Wu, Z.; Ramsundar, B.; Feinberg, E. N.; Gomes, J.; Geniesse, C.; Pappu, A. S.; Leswing, K.; and Pande, V. 2018. MoleculeNet: a benchmark for molecular machine learning. Chemical science, 9(2): 513--530

    2018

  38. [46]

    Xu, K.; Hu, W.; Leskovec, J.; and Jegelka, S. 2019. How Powerful are Graph Neural Networks? In Proceedings of the 7th International Conference on Learning Representations ( ICLR 2019) . New Orleans, LA, USA: OpenReview.net

    2019

  39. [47]

    Xu, M.; Wang, H.; Ni, B.; Guo, H.; and Tang, J. 2021. Self-supervised Graph-level Representation Learning with Local and Global Structure. In Proceedings of the 38th International Conference on Machine Learning ( ICML 2021) , volume 139 of PMLR, 11548--11558. Virtual Event: PMLR

    2021

  40. [48]

    You, Y.; Chen, T.; Shen, Y.; and Wang, Z. 2021. Graph Contrastive Learning Automated. In Proceedings of the 38th International Conference on Machine Learning ( ICML 2021) , volume 139 of PMLR, 12121--12132. Virtual Event: PMLR

    2021

  41. [49]

    You, Y.; Chen, T.; Sui, Y.; Chen, T.; Wang, Z.; and Shen, Y. 2020. Graph Contrastive Learning with Augmentations. In Proceedings of the 33rd Annual Conference on Neural Information Processing Systems (NeurIPS 2020). Virtual Event

    2020

  42. [50]

    Yu, J.; Cao, J.; and He, R. 2022. Improving Subgraph Recognition with Variational Graph Information Bottleneck. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition ( CVPR 2022) , 19374--19383. New Orleans, LA, USA: IEEE

    2022

  43. [51]

    Yu, J.; Xu, T.; Rong, Y.; Bian, Y.; Huang, J.; and He, R. 2021. Graph Information Bottleneck for Subgraph Recognition. In Proceedings of the 9th International Conference on Learning Representations ( ICLR 2021) . Virtual Event: OpenReview.net

    2021

  44. [52]

    Zhang, J.; Zhang, H.; Xia, C.; and Sun, L. 2020. Graph-Bert: Only Attention is Needed for Learning Graph Representations. CoRR, abs/2001.05140

    2020 arXiv

  45. [53]

    Zhang, Z.; Liu, Q.; Wang, H.; Lu, C.; and Lee, C. 2021. Motif-based Graph Self-Supervised Learning for Molecular Property Prediction. In Proceedings of the 34th Advances in Neural Information Processing Systems (NeurIPS 2021), 15870--15882. Virtual Event

    2021

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.