Pith. sign in

REVIEW 3 major objections 4 minor 67 references

LatentFlow: Visual Analytics for Latent Space Analysis in Molecular Graph Neural Networks

T0 review · 3 major / 4 minor · reviewed 2026-08-01 · deepseek-v4-flash

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

Pith's one-line read LatentFlow lets chemists watch molecular clusters evolve through GNN layers and training.

desk verdict The two-dimensional Sankey layout is a real contribution; the evaluation is too thin to support the abstract's claim about helping scientists. read the letter →

arxiv 2607.21941 v1 pith:LECNRLR3 submitted 2026-07-24 cs.LG cs.HC

Shiyi Liu , Jiaqing Chen , Nicholas Hadler , Rostyslav Hnatyshyn , Michael W. Mahoney , Talita Perciano , John F. Hartwig , Gunther H. Weber
show 1 more author
Ross Maciejewski
This is my paper · ORCID
classification cs.LGcs.HC
keywords latentspaceanalysisgraphneuralnetworksvisualanalyticsmolecularpropertypredictionclustertransitionsSankeydiagramembeddinginterpretationcheminformatics
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

The paper claims that a new visual analytics system, LatentFlow, lets computational chemists see how a molecular GNN reorganizes its internal representations across layers and training states. It clusters embeddings at each stage, connects the clusters with flows in a modified Sankey diagram, and links them to molecular substructures, so scientists can watch clusters split, merge, or stabilize and interpret what chemistry drives the changes. In two expert-led case studies the system revealed a shift from taxonomy-aligned to property-aligned organization and exposed differences between augmented and baseline training. If correct, LatentFlow is a diagnostic tool for understanding and steering embedding-based molecular models.

What carries the argument

The modified Sankey diagram: a matrix of cells, each holding cluster rectangles, with red and green flows along two axes and gap heatmaps showing the fraction of molecules that change cluster between neighboring cells. It is the mechanism that makes multi-dimensional evolution visible at a glance. Around it, clustering methods group embeddings, and cluster members are linked to molecular structures via common-substructure extraction and fingerprint-based similarity, tying every flow to concrete chemistry.

What would settle it

Run a negative-control training run with shuffled solubility labels; if LatentFlow still shows late-epoch clusters arranged along a solubility gradient, then the claimed 'property-driven organization' is an artifact of projection or clustering rather than a learned signal. Separately, have a fresh group of chemists who never saw LatentFlow reproduce the two case-study findings; failure to do so would undercut the generality claim.

Watch

Extended reading notes

Core claim

The paper claims that a two-dimensional flow visualization—clusters of molecular embeddings shown as rectangles in a grid, with red horizontal flows and green vertical flows carrying molecules between clusters—lets experts see how a GNN's latent space reorganizes. In one demonstration, early-training clusters matched a coarse aryl/alkyl taxonomy, then drifted apart as the model learned to organize molecules by solubility; in another, a model trained with multiple conformations per molecule produced clusters aligned with selectivity while a single-conformation model did not. The authors take these cases as evidence the system surfaces chemically meaningful patterns.

Load-bearing premise

The evaluation rests on a single domain expert who co-designed the system; if his fluent use reflects co-design familiarity rather than general usability, the evidence that LatentFlow helps scientists broadly is not established.

Editorial extensions

If this is right

  • Experts can locate the exact layer and epoch at which a model abandons one organizing scheme for another.
  • Augmentation strategies can be compared by whether their latent spaces show target-aligned gradients.
  • The same cluster-flow design transfers to any embedding space that changes along two dimensions, such as temporal drift or domain adaptation.
  • The Top-k and cluster-member panels allow detection of substructure-driven grouping that aggregate metrics would miss.

Reading between the lines

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

  • A training-time monitor could flag when the gap heatmap stabilizes, saving chemists from inspecting every epoch.
  • The same two-axis flow logic could diagnose catastrophic forgetting, where clusters split and merge across tasks and time.
  • The observed solubility alignment could be confirmed with a negative control: shuffle labels and see if the drift persists.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. LatentFlow is a visual analytics system for exploring how molecular GNN latent spaces evolve across network layers and model states. The system clusters molecule embeddings, visualizes cluster transitions with a two-dimensional modified Sankey diagram, and links clusters to molecular structures, substructures, Tanimoto similarity, model predictions, and expert-uploaded labels. The paper reports two case studies: (1) a Chemprop model on ESOL where clusters shift from an aryl/alkyl taxonomy to solubility-related organization, and (2) two DimeNet models with and without conformer augmentation, where the augmented model shows target-aligned cluster structure and different treatment of a phosphite substructure. The central claim is that LatentFlow helps scientists understand latent-space evolution, identify meaningful molecular patterns, and interpret model behavior.

Significance. If supported, LatentFlow would address a genuine gap: existing latent-space visualizations mostly inspect a single state or pairwise comparisons, whereas LatentFlow explicitly supports tracking cluster transitions across two dimensions such as layer and epoch. The modified Sankey layout is a credible design contribution, and the system is described in enough detail to be reproduced; an online demo is provided. The two case studies are internally coherent and the chemical findings are plausible. However, the evidence for the central effectiveness claim is currently weak: the only human evaluation is conducted by the same domain expert who co-designed the system over five months, with no independent users, no baseline comparison, and no quantitative outcome measures. The paper itself concedes in Sec. 7.2 that the needs are not claimed to be representative and in Sec. 7.3 that the insights depend heavily on clustering/projection settings. These caveats bound what the paper can claim but do not repair the gap between the abstract's claim and the evidence provided.

major comments (3)
  1. [Sec. 4 and Sec. 6] The only effectiveness evidence is two case studies conducted with the same domain expert who participated in the five-month iterative design process (Sec. 4 and Sec. 6). This cannot, by itself, establish the abstract's claim that 'LatentFlow helps scientists understand how latent spaces evolve, identify meaningful molecular patterns, and better interpret model behavior.' The co-designing expert is not an independent user: the expert's familiarity with the intended workflow and with the system's interaction model may explain the reported findings. The paper needs either an evaluation with independent domain experts (with a defined protocol and task-based measures such as number of verified insights, accuracy, or time), or a comparative study against a baseline tool (e.g., static t-SNE/UMAP scatterplots plus RDKit or DataWarrior) to support the generalization implied by the abstract. The
  2. [Sec. 7.3 and Sec. 5.1] The paper states in Sec. 7.3 that insights from the system 'are highly dependent on choosing suitable clustering and projections,' yet the case studies do not test robustness to these choices. For instance, the aryl/alkyl-to-solubility transition in Case Study 1 and the substructure-grouping difference in Case Study 2 are reported for a particular clustering method and set of parameters, but no sensitivity analysis is shown. The system exposes clustering and projection controls (Sec. 5.1), so the authors already have the infrastructure to sweep settings; they should report whether the central observations—AMI trends, cluster transitions, and substructure enrichments—are stable across reasonable choices of clustering method, cluster count, and projection. Without such a check, the observed patterns may be artifacts of a single configuration, and the paper's own Sec. 7.3 limitation becomes
  3. [Sec. 6.1 and Sec. 6.2] Several load-bearing claims in the case studies could and should be quantified with metrics that the system itself already computes or could easily compute. For example, 'Cluster 0 contains methane-like substructures' and 'Cluster 1 consistently contains halogen atoms' could be supported by substructure enrichment relative to the full dataset; the claim that label agreement weakens across epochs could be supported by reporting AMI values; and the claim that the 10conf model's clusters are 'better aligned with target values' could be supported by silhouette scores or target-value separation within clusters. The qualitative narrative is suggestive, but it does not provide evidence that the observed patterns are statistically meaningful or reproducible. Reporting these numbers would strengthen the central argument without changing the system.
minor comments (4)
  1. [Sec. 6.2] The text uses 'Dimenet' once; the architecture is spelled 'DimeNet' elsewhere. Also, 'Fig. Fig. 1G' in Sec. 5.1 contains a duplicated 'Fig.'.
  2. [References] References [19] and [20] appear to be the same paper (Gómez-Bombarelli et al., ACS Central Science 2018). Reference [20] is cited in Sec. 6.2 for the claim that conformer augmentation improved extrapolation, which seems more likely to refer to the accompanying DimeNet study [21] or a related augmentation experiment. Please check and correct the citation.
  3. [Sec. 5.2 and Fig. 1] The figure callouts A1, A2, A3 are used in the caption but are not fully explained in the text. In particular, the 'gap heatmaps' (A2) and the relationship between horizontal and vertical gap color encoding should be described more explicitly so a reader can map the described encoding to the figure.
  4. [Sec. 5.6] The third column of the Top-k Panel is described as 'Fig. 1E3' in the text, but the figure appears to label representative molecules as E4. Please correct the mislabeled reference.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: LatentFlow is a systems paper with no derivation-based predictions; the self-referential evaluation is a validity limitation, not a circular step.

full rationale

LatentFlow makes no first-principles derivation or prediction claim; it presents a visual analytics workflow, so there is no fitted parameter renamed as a prediction and no equation-level reduction of output to input. The clustering and projection steps (Sec. 5) are user-configurable choices, and the paper explicitly acknowledges in Sec. 7.3 that "the insights gleaned from the system are highly dependent on choosing suitable clustering and projections" — a robustness limitation, not a circularity. Sec. 7.2 also concedes that "our study does not claim that these needs are representative of all molecular machine learning practitioners"; again a generalization limitation. The effectiveness evidence in Sec. 6 comes from two case studies "conducted with the same domain expert who participated in the design process" (Sec. 6), which is a genuine threat to external validity but not a logical circularity: the expert's qualitative observations do not by construction entail the conclusion that the system helps scientists. No load-bearing self-citation is used: overlapping-author prior work (e.g., [21] for the DimeNet models) supplies model context, not the justification for the system's claims. Therefore no circular step is exhibited.

Assumptions & free parameters 3 free parameters · 3 assumptions · 0 invented entities

The system introduces no new physical entities or mathematical axioms. The load-bearing assumptions are domain-level: that embeddings cluster meaningfully, that projections preserve enough structure, and that substructure fingerprints capture the chemistry relevant to the clusters. The main free parameters are interactive analysis choices (clustering, top-k, projection) that the paper itself says can change the insights (Sec. 7.3).

free parameters (3)
  • Cluster count / clustering parameters per latent state = not reported (typically ≤10)
    LatentFlow groups embeddings using k-NN, AHC, HDBSCAN, or PG-Means (Sec. 5); the number of clusters and parameters are user choices that determine all Sankey transitions and case-study patterns.
  • Top-k representative molecules = not reported (user-adjustable)
    Top-k Panel (Sec. 5.6) summarizes each cluster using the k molecules closest to the centroid; k is selected by the user and changes which substructures are extracted.
  • Projection method and temporally consistent layout parameters = not reported
    Scatterplot View uses PCA, t-SNE, or UMAP with dynamic layouts (Sec. 5.4); the choice affects apparent cluster separation used in both case studies.
assumptions (3)
  • domain assumption Molecular GNN latent spaces admit discrete clusters whose transitions across layers and epochs reflect meaningful model organization
    Sec. 3.1 and DR1; the Modified Sankey View renders clusters as rectangles and flows, so the entire analysis presupposes clustering is a valid summarization.
  • domain assumption Temporally consistent low-dimensional projections preserve cross-state comparability
    Sec. 5.4 uses shared PCA bases and dynamic t-SNE/UMAP; paper acknowledges distortions in Sec. 7.3.
  • domain assumption Tanimoto similarity and common-substructure matching adequately summarize cluster chemistry
    Sec. 5.6 uses fingerprints and shared subgraphs to characterize clusters; case-study interpretations rely on these summaries.

how reviews work

0 comments
Cite this review

Pith. "Pith review of LatentFlow: Visual Analytics for Latent Space Analysis in Molecular Graph Neural Networks." pith.science (2026). https://pith.science/paper/LECNRLR3

@misc{pith2026260721941,
  author       = {Pith},
  title        = {Pith review of: LatentFlow: Visual Analytics for Latent Space Analysis in Molecular Graph Neural Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LECNRLR3}},
  note         = {Machine review of arXiv:2607.21941}
}
read the original abstract

Chemists and materials scientists increasingly use machine learning models, such as graph neural networks (GNNs), to predict properties of molecules and the outcomes of their reactions. Beyond predictive performance, understanding how these models organize chemical information internally in their latent spaces, i.e., the embeddings of the molecules, is critical. Analyzing latent spaces helps diagnose model behavior and assess whether the learned embeddings are organized in ways that reflect meaningful chemical relationships. Unfortunately, existing methods provide limited support for analyzing latent spaces across layers and across different model states (e.g., training epochs, model configurations, and input data), making it difficult to understand how these latent spaces evolve throughout a model or relate to chemical concepts. We present LatentFlow, a visual analytics system developed in collaboration with a domain expert for analyzing latent spaces in molecular GNNs. LatentFlow groups embeddings into clusters and supports exploration of latent spaces by tracking how these clusters change across layers and model states using a modified Sankey diagram. To support interpretation, LatentFlow links these clusters to representative molecules and their shared substructures, and it allows scientists to introduce their own domain knowledge and compare it with the patterns found in the latent spaces. We evaluate LatentFlow through two case studies. The results show that LatentFlow helps scientists understand how latent spaces evolve, identify meaningful molecular patterns, and better interpret model behavior.

Figures

Figures reproduced from arXiv: 2607.21941 by the authors.

Figure 1
Figure 1. LatentFlow interface for analyzing latent spaces in molecular GNNs through coordinated views. (A) The Modified Sankey View shows cluster changes across two model-state dimensions. (B) The horizontal (B1) and vertical Sankey View (B2) reveal detailed transitions along a single dimension. (C) The scatterplot projects embeddings into 2D to inspect how molecules are distributed and separated. (D) The cluster member view… view at source ↗
Figure 2
Figure 2. LatentFlow integrates model data and domain knowledge into a four-level visual analytics system. (A) Model data, including molecular structures, embeddings extracted from different layers and model states, and model predictions, are provided as input. (B) At the overview level, embeddings are grouped into clusters using clustering methods and visualized in the Modified Sankey View to show how molecules move between … view at source ↗
Figure 3
Figure 3. Several designs for the Sankey View; C is employed in [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Case study 1: analysis of a Chemprop model showing the transition from taxonomy-aligned to property-driven latent space organization. [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Case study 1: Cluster 0 at Epoch 19/gnn3 (A) The Top-k Panel shows representative molecules and shared substructures. (B) The Cluster Member Panel shows individual molecules in the cluster. the Modified Sankey View, as shown in Fig. 4A. The cells are shown with a blue …
Figure 7
Figure 7. Figure 7: Case study 2: comparison of two DimeNet models with the same architecture but different training inputs. (A) The Modified Sankey View [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

67 extracted references · 10 canonical work pages

  1. [20]

    Gómez-Bombarelli, J

    R. Gómez-Bombarelli, J. N. Wei, D. Duvenaud, J. M. Hernández-Lobato, B. Sánchez-Lengeling, D. Sheberla et al. Automatic chemical design using a data-driven continuous representation of molecules.ACS Central Science, 4(2):268–276, 2018. PMID: 29532027. doi: 10.1021/acscentsci. 7b00572 8

    doi:10.1021/acscentsci 2018

  2. [1]

    Achille and S

    A. Achille and S. Soatto. Emergence of invariance and disentanglement in deep representations.Journal of Machine Learning Research, 19(50):1–34,

  3. [2]

    Andronov, M

    M. Andronov, M. V . Fedorov, and S. Sosnin. Exploring chemical reaction space with reaction difference fingerprints and parametric t-sne.ACS Omega, 6(45):30743–30751, 2021. PMID: 34805702. doi: 10.1021/ acsomega.1c04778 2

    2021

  4. [3]

    K. Atz, F. Grisoni, and G. Schneider. Geometric deep learning on molecu- lar representations.Nature Machine Intelligence, 3(12):1023–1032, 2021. doi: 10.1038/s42256-021-00418-8 2

    doi:10.1038/s42256-021-00418-8 2021

  5. [4]

    Bajusz, A

    D. Bajusz, A. Rácz, and K. Héberger. Why is tanimoto index an appro- priate choice for fingerprint-based similarity calculations?Journal of Cheminformatics, 7(1):20, 2015. doi: 10.1186/s13321-015-0069-3 3

    doi:10.1186/s13321-015-0069-3 2015

  6. [5]

    Bengio, A

    Y . Bengio, A. Courville, and P. Vincent. Representation learning: A review and new perspectives.IEEE Transactions on Pattern Analysis and Machine Intelligence, 35(8):1798–1828, 2013. doi: 10.1109/TPAMI.2013. 50 2

    work page doi:10.1109/tpami.2013 2013

  7. [6]

    Bienfait and P

    B. Bienfait and P. Ertl. Jsme: a free molecule editor in javascript.Journal of cheminformatics, 5(1):24, 2013. doi: 10.1186/1758-2946-5-24 6

    work page doi:10.1186/1758-2946-5-24 2013

  8. [7]

    J. W. Burns, A. S. Zalte, C. R. A. Abreu, J. Sieg, C. Feldmann, M. Mathea et al. Deep Learning Foundation Models from Classical Molecular De- scriptors, Feb. 2026. doi: 10.48550/arXiv.2506.15792 2

    doi:10.48550/arxiv.2506.15792 2026

Show all 67 references
  1. [8]

    Bushuiev, A

    R. Bushuiev, A. Bushuiev, R. Samusevich, C. Brungs, J. Sivic, and T. Pluskal. Self-supervised learning of molecular representations from millions of tandem mass spectra using dreams.Nature Biotechnology,

  2. [9]

    R. J. G. B. Campello, D. Moulavi, and J. Sander. Density-based clustering based on hierarchical density estimates. In J. Pei, V . S. Tseng, L. Cao, H. Motoda, and G. Xu, eds.,Advances in Knowledge Discovery and Data Mining, pp. 160–172. Springer Berlin Heidelberg, Berlin, Heidelberg,

  3. [10]

    S. Chen, A. Wulamu, Q. Zou, H. Zheng, L. Wen, X. Guo et al. Md-gnn: A mechanism-data-driven graph neural network for molecular properties prediction and new material discovery.Journal of Molecular Graphics and Modelling, 123:108506, 2023. doi: 10.1016/j.jmgm.2023.108506 2

    2023

  4. [11]

    Cilibrasi and P

    R. Cilibrasi and P. Vitanyi. Clustering by compression.IEEE Transactions on Information Theory, 51(4):1523–1545, 2005. doi: 10.1109/TIT.2005. 844059 4

    2005 doi

  5. [12]

    Cover and P

    T. Cover and P. Hart. Nearest neighbor pattern classification.IEEE Transactions on Information Theory, 13(1):21–27, 1967. doi: 10.1109/ TIT.1967.1053964 4

    1967

  6. [13]

    Faust, D

    R. Faust, D. Glickenstein, and C. Scheidegger. Dimreader: Axis lines that explain non-linear projections.IEEE Transactions on Visualization and Computer Graphics, 25(1):481–490, 2019. doi: 10.1109/TVCG.2018. 2865194 2

    2019 doi

  7. [14]

    Feng and G

    Y . Feng and G. Hamerly. Pg-means: learning the number of clusters in data. In B. Schölkopf, J. Platt, and T. Hoffman, eds.,Advances in Neural Information Processing Systems, vol. 19. MIT Press, 2006. doi: 10. 7551/mitpress/7503.003.0054 4

    2006

  8. [15]

    E. B. Fowlkes and C. L. Mallows. A method for comparing two hi- erarchical clusterings.Journal of the American Statistical Association, 78(383):553–569, 1983. doi: 10.1080/01621459.1983.10478008 4

    1983

  9. [16]

    Gasteiger, J

    J. Gasteiger, J. Groß, and S. Günnemann. Directional message passing for molecular graphs. InInternational Conference on Learning Representa- tions, 2020. 2

    2020

  10. [17]

    González-Medina and J

    M. González-Medina and J. L. Medina-Franco. Platform for unified molecular analysis: Puma.Journal of Chemical Information and Model- ing, 57(8):1735–1740, 2017. PMID: 28737911. doi: 10.1021/acs.jcim. 7b00253 3

    2017 doi

  11. [18]

    Goodfellow, Y

    I. Goodfellow, Y . Bengio, A. Courville, and Y . Bengio.Deep learning, vol. 1. MIT press Cambridge, 2016. doi: 10.4258/hir.2016.22.4.351 2

    2016 doi

  12. [21]

    Hadler, N

    N. Hadler, N. I. Rinehart, M. Elkin, J. Nicolai, G. Gheibi, J. Chen et al. A 3d, structure-based, deep learning approach for predicting the regioselectivity of transition-metal catalysis.ChemRxiv, 2026(0128),

    2026

  13. [22]

    Harrower and C

    M. Harrower and C. A. Brewer. Colorbrewer. org: an online tool for selecting colour schemes for maps.The Cartographic Journal, 40(1):27– 37, 2003. doi: 10.1179/000870403235002042 5

    2003 doi

  14. [23]

    E. Heid, K. P. Greenman, Y . Chung, S.-C. Li, D. E. Graff, F. H. Vermeire et al. Chemprop: A machine learning package for chemical property prediction.Journal of Chemical Information and Modeling, 64(1):9–17,

  15. [24]

    T. Hua, W. Wang, Z. Xue, S. Ren, Y . Wang, and H. Zhao. On feature decorrelation in self-supervised learning. InProceedings of the IEEE/CVF international conference on computer vision, pp. 9598–9608, 2021. 2

    2021

  16. [25]

    Huang, D

    Z. Huang, D. Witschard, K. Kucher, and A. Kerren. Va + embeddings star: A state-of-the-art report on the use of embeddings in visual analytics. Computer Graphics Forum, 42(3):539–571, 2023. doi: 10.1111/cgf.14859 2

    2023 doi

  17. [26]

    Humer, H

    C. Humer, H. Heberle, F. Montanari, T. Wolf, F. Huber, R. Henderson et al. Cheminformatics model explorer (cime): exploratory analysis of chemical model explanations.Journal of Cheminformatics, 14(1):21, 2022. doi: 10. 1186/s13321-022-00600-z 3

    2022

  18. [27]

    Jiménez-Luna, F

    J. Jiménez-Luna, F. Grisoni, and G. Schneider. Drug discovery with explainable artificial intelligence.Nature Machine Intelligence, 2(10):573– 584, 2020. doi: 10.1038/s42256-020-00236-4 2

    2020 doi

  19. [28]

    S. C. Johnson. Hierarchical clustering schemes.Psychometrika, 32(3):241–254, 1967. doi: 10.1007/BF02289588 4

    1967 doi

  20. [29]

    J. H. W. Jr. Hierarchical grouping to optimize an objective function. Journal of the American Statistical Association, 58(301):236–244, 1963. doi: 10.1080/01621459.1963.10500845 4

    1963

  21. [30]

    B. Kale, A. Clyde, M. Sun, A. Ramanathan, R. Stevens, and M. E. Papka. Chemograph: Interactive visual exploration of the chemical space.Com- puter Graphics Forum, 42(3):13–24, 2023. doi: 10.1111/cgf.14807 3

    2023 doi

  22. [31]

    D. Keim, G. Andrienko, J.-D. Fekete, C. Görg, J. Kohlhammer, and G. Melançon.Visual Analytics: Definition, Process, and Challenges, pp. 154–175. Springer Berlin Heidelberg, Berlin, Heidelberg, 2008. doi: 10. 1007/978-3-540-70956-5_7 9

    2008

  23. [32]

    Kornblith, M

    S. Kornblith, M. Norouzi, H. Lee, and G. Hinton. Similarity of neural network representations revisited. In K. Chaudhuri and R. Salakhutdinov, eds.,Proceedings of the 36th International Conference on Machine Learn- ing, vol. 97 ofProceedings of Machine Learning Research, pp. 3...

    2019

  24. [33]

    M. J. Kusner, B. Paige, and J. M. Hernández-Lobato. Grammar variational autoencoder. In D. Precup and Y . W. Teh, eds.,Proceedings of the 34th International Conference on Machine Learning, vol. 70 ofProceedings of Machine Learning Research, pp. 1945–1954. PMLR, 06–11 Aug 2017. 3

    1945

  25. [34]

    G. Landrum. Rdkit: Open-source cheminformatics software. https: //www.rdkit.org, 2016. Accessed: 2026-03-12. 4

    2016

  26. [35]

    A. C. Li, A. A. Efros, and D. Pathak. Understanding collapse in non- contrastive siamese representation learning. In S. Avidan, G. Brostow, M. Cissé, G. M. Farinella, and T. Hassner, eds.,Computer Vision – ECCV 2022, pp. 490–505. Springer Nature Switzerland, Cham, 2022. doi: 1...

    2022

  27. [36]

    Q. Li, K. S. Njotoprawiro, H. Haleem, Q. Chen, C. Yi, and X. Ma. Embed- dingvis: A visual analytics approach to comparative network embedding inspection. In2018 IEEE Conference on Visual Analytics Science and Technology (VAST), pp. 48–59, 2018. doi: 10.1109/V AST.2018.8802454 2

    2018

  28. [37]

    Y . Liu, E. Jun, Q. Li, and J. Heer. Latent space cartography: Visual analysis of vector space embeddings.Computer Graphics Forum, 38(3):67–78,

  29. [38]

    Z. Liu, Y . Wang, J. Bernard, and T. Munzner. Visualizing graph neural networks with corgie: Corresponding a graph to its embedding.IEEE Transactions on Visualization and Computer Graphics, 28(6):2500–2516,

  30. [39]

    Ma´ckiewicz and W

    A. Ma´ckiewicz and W. Ratajczak. Principal components analysis (pca). Computers & Geosciences, 19(3):303–342, 1993. doi: 10.1016/0098-3004 (93)90090-R 2

    1993 doi

  31. [40]

    McInnes, J

    L. McInnes, J. Healy, and J. Melville. Umap: Uniform manifold approx- imation and projection for dimension reduction, 2020. doi: arXiv.1802. 03426 2

    2020

  32. [41]

    M. Meil˘a. Comparing clusterings—an information based distance.Journal of Multivariate Analysis, 98(5):873–895, 2007. doi: 10.1016/j.jmva.2006. 11.013 4

    2007 doi

  33. [42]

    Y . Miao, H. Ma, and J. Huang. Recent advances in toxicity prediction: Applications of deep graph learning.Chemical Research in Toxicology, 36(8):1206–1226, 2023. PMID: 37562046. doi: 10.1021/acs.chemrestox. 2c00384 2

    2023 doi

  34. [43]

    Monem, A

    S. Monem, A. H. Abdel-Hamid, and A. E. Hassanien. Drug toxicity prediction model based on enhanced graph neural network.Computers in Biology and Medicine, 185:109614, 2025. doi: 10.1016/j.compbiomed. 2024.109614 2

    2025

  35. [44]

    Probst and J.-L

    D. Probst and J.-L. Reymond. Visualization of very large high-dimensional data sets as minimum spanning trees.Journal of Cheminformatics, 12(1):12, 2020. doi: 10.1186/s13321-020-0416-x 3

    2020 doi

  36. [45]

    Y . A. Qadri, S. Shaikh, K. Ahmad, I. Choi, S. W. Kim, and A. V . Vasi- lakos. Explainable artificial intelligence: A perspective on drug discovery. Pharmaceutics, 17(9):1119, 2025. doi: 10.3390/pharmaceutics17091119 2

    2025 doi

  37. [46]

    Raghu, J

    M. Raghu, J. Gilmer, J. Yosinski, and J. Sohl-Dickstein. Svcca: Singular vector canonical correlation analysis for deep learning dynamics and interpretability.Advances in neural information processing systems, 30,

  38. [47]

    P. E. Rauber, S. G. Fadel, A. X. Falcão, and A. C. Telea. Visualizing the hidden activity of artificial neural networks.IEEE Transactions on Visualization and Computer Graphics, 23(1):101–110, 2017. doi: 10. 1109/TVCG.2016.2598838 2

    2017

  39. [48]

    P. E. Rauber, A. X. Falcão, and A. C. Telea. Visualizing time-dependent data using dynamic t-sne. InProceedings of the Eurographics / IEEE VGTC Conference on Visualization: Short Papers, EuroVis ’16, pp. 73–77. Eurographics Association, Goslar, DEU, 2016. doi: 10.1111/cgf.12878 4

    2016 doi

  40. [49]

    Reiser, M

    P. Reiser, M. Neubert, A. Eberhard, L. Torresi, C. Zhou, C. Shao et al. Graph neural networks for materials science and chemistry.Commu- nications Materials, 3(1):93, 2022. doi: 10.1038/s43246-022-00315-6 2

    2022 doi

  41. [50]

    M. V . Sabando, P. Ulbrich, M. Selzer, J. Byška, J. Miˇcan, I. Ponzoni et al. ChemV A: Interactive visual analysis of chemical compound similarity in virtual screening.IEEE Transactions on Visualization and Computer Graphics, 27(2):891–901, 2021. doi: 10.1109/TVCG.2020.3030438 3

    2021

  42. [51]

    Sander, J

    T. Sander, J. Freyss, M. von Korff, and C. Rufener. Datawarrior: An open-source program for chemistry aware data visualization and analysis. Journal of Chemical Information and Modeling, 55(2):460–473, 2015. PMID: 25558886. doi: 10.1021/ci500588j 3

    2015 doi

  43. [52]

    Schäfer, N

    M. Schäfer, N. Brich, J. Byška, S. M. Marques, D. Bedná ˇr, P. Thiel et al. Invado: Interactive visual analysis of molecular docking data.IEEE Transactions on Visualization and Computer Graphics, 30(4):1984–1997,

    1984

  44. [53]

    X. Shi, L. Zhou, Y . Huang, Y . Wu, and Z. Hong. A review on the ap- plications of graph neural networks in materials science at the atomic scale.Materials Genome Engineering Advances, 2(2):e50, 2024. doi: 10. 1002/mgea.50 2

    2024

  45. [54]

    Shilpa, G

    S. Shilpa, G. Kashyap, and R. B. Sunoj. Recent applications of machine learning in molecular property and chemical reaction outcome predictions. The Journal of Physical Chemistry A, 127(40):8253–8271, 2023. PMID: 37769193. doi: 10.1021/acs.jpca.3c04779 2

    2023 doi

  46. [55]

    Smilkov, N

    D. Smilkov, N. Thorat, C. Nicholson, E. Reif, F. B. Viégas, and M. Wat- tenberg. Embedding projector: Interactive visualization and interpretation of embeddings, 2016. doi: 10.48550/arXiv.1611.05469 2

    2016 doi arXiv

  47. [56]

    Stahnke, M

    J. Stahnke, M. Dörk, B. Müller, and A. Thom. Probing projections: Interac- tion techniques for interpreting arrangements and errors of dimensionality reductions.IEEE Transactions on Visualization and Computer Graphics, 22(1):629–638, 2016. doi: 10.1109/TVCG.2015.2467717 2

    2016

  48. [57]

    Van der Maaten and G

    L. Van der Maaten and G. Hinton. Visualizing data using t-sne.Journal of machine learning research, 9(11), 2008. 2

    2008

  49. [58]

    Z. J. Wang, F. Hohman, and D. H. Chau. WizMap: Scalable interac- tive visualization for exploring large machine learning embeddings. In D. Bollegala, R. Huang, and A. Ritter, eds.,Proceedings of the 61st An- nual Meeting of the Association for Computational Linguistics (Volume...

    2023 doi

  50. [59]

    doi: 10.1109/TVCG.2023.3337642 3

    2023

  51. [60]

    Z. Wu, B. Ramsundar, E. N. Feinberg, J. Gomes, C. Geniesse, A. S. Pappu et al. Moleculenet: a benchmark for molecular machine learning.Chemical science, 9(2):513—530, January 2018. doi: 10.1039/c7sc02664a 6

    2018 doi

  52. [61]

    Y . Zou, B. Li, and R. Ran. Joint umap for visualization of time-dependent data. InComputational Visual Media: 13th International Conference, CVM 2025, Hong Kong SAR, China, April 19–21, 2025, Proceedings, Part III, pp. 266–288. Springer-Verlag, Berlin, Heidelberg, 2025. doi: ...

    2025

  53. [66]

    Winter, F

    R. Winter, F. Montanari, F. Noé, and D.-A. Clevert. Learning continuous and data-driven molecular descriptors by translating equivalent chem- ical representations.Chem. Sci., 10:1692–1701, 2019. doi: 10.1039/ C8SC04175J 3

    2019

  54. [2013]

    doi: 10.1007/978-3-642-37456-2_14 4

    doi

  55. [2019]

    doi: 10.1111/cgf.13672 2

    doi

  56. [2022]

    doi: 10.1109/TVCG.2022.3148197 2

    2022

  57. [2024]

    doi: 10.1021/acs.jcim.3c01250 2, 6

    PMID: 38147829. doi: 10.1021/acs.jcim.3c01250 2, 6

    doi

  58. [2025]

    doi: 10.1038/s41587-025-02663-3 2

    doi

  59. [2026]

    doi: 10.26434/chemrxiv.10001648/v1 6

    doi

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.