Pith. sign in

REVIEW 3 major objections 5 minor 34 references

Can Molecular Evolution Mechanism Enhance Molecular Representation?

T0 review · 3 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read Tracing a molecule's evolutionary path cuts property-prediction error by 32%.

desk verdict The MEvoN idea is worth a look, but the 32.3% gain is not established: EvoL consumes labels along full-dataset paths, and the ablation shows the path-only module actually hurts. read the letter →

arxiv 2501.15799 v1 pith:FXVOU43F submitted 2025-01-27 q-bio.BM cs.LGcs.NE

Kun Li , Longtao Hu , Xiantao Cai , Jia Wu , Wenbin Hu This is my paper · ORCID
classification q-bio.BMcs.LGcs.NE
keywords molecularrepresentationevolutionarynetworkpropertypredictionQM9QM7graphneuralnetworkssimilarityHOMO-LUMOgap
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 adding a molecule's evolutionary history as auxiliary input to a standard molecular encoder significantly improves property prediction. The authors construct a molecular evolutionary network (MEvoN) by grouping molecules by atom count and linking similar molecules across groups, then extract structural path and property-label sequences from this network. Plugging these sequences into GCN, GIN, SchNet, and ComENet yields a reported average 32.3% reduction in mean absolute error on the QM7 and QM9 benchmark tasks. If this is right, representation learning for molecules gains from relational structure among similar molecules, not just from each molecule's isolated structure. The evaluation, however, assumes that property labels of ancestor molecules along each test molecule's path are available as input.

What carries the argument

The load-bearing object is the molecular evolutionary network (MEvoN): a directed network in which molecules are nodes grouped by atom count, and edges point from smaller 'ancestor' molecules to larger 'descendant' molecules when edit-distance or fingerprint Tanimoto similarity exceeds a threshold $ heta_1$ and Weisfeiler-Lehman graph-kernel similarity exceeds a second threshold $ heta_2$. From this network, evolutionary paths are extracted by backtracking from a target molecule to the root, and both the sequence of molecular embeddings and the sequence of property labels along each path are fed through Transformer encoders. The resulting path and label features are concatenated with the base encoder's own features to produce the final prediction, which is what carries the argument that evolutionary context enhances representation.

What would settle it

Re-run the QM9 experiments after building the MEvoN exclusively from the training split, so that every test molecule's evolutionary path lies entirely inside the training set; alternatively, remove the EvoL label-input module and keep only structural path features. If the reported 32.3% average MAE gain collapses under either protocol, the improvement depends on transductive label access rather than on evolutionary representation.

Watch

Extended reading notes

Core claim

The central claim is that a molecule's evolutionary trajectory, not only its static structure, carries predictive signal for quantum-chemical properties. MEvoN-MPP combines three components: a molecular encoder (MolE) that can be any existing model; an EvoP module that encodes the sequence of graph features along the evolutionary path with a Transformer; and an EvoL module that encodes the property labels along that same path. The model simultaneously predicts the target molecule's property and the property difference between the last two molecules on the path, masking the target's own label. On QM9 the method reports an average MAE improvement of 32.3% across the four encoders, with the largest gains on GCN (53.11%) and GIN (35.00%), and on QM7 a GCN-based variant reaches MAE 45.9, competitive with much larger pretrained models.

Load-bearing premise

The evaluation assumes that, for every test molecule, the property labels of the smaller molecules along its evolutionary path are available as input, with only the test molecule's own label masked.

Editorial extensions

If this is right

  • Because MolE can be any molecule encoder, the reported gains, if correct, transfer across GNNs, CNNs, and Transformers without architectural changes to the base model.
  • The SHAP analysis of mutation types suggests the network captures chemically interpretable regularities: adding rings or oxygen/nitrogen atoms tends to lower the HOMO-LUMO gap, while adding a three-membered ring raises it.
  • The hyperparameter study indicates that using 3 to 5 evolutionary paths per molecule gives stable predictions; fewer paths lose diversity and more paths add noise and computation.
  • On QM7, MEvoN-MPP with a GCN encoder reaches MAE 45.9 without pretraining, placing it in the range of methods that use large-scale pretraining and fine-tuning.

Reading between the lines

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

  • If the ancestor labels are not strictly restricted to the training split, the 32.3% gain may be partly a transductive label-propagation effect; the fair test would rebuild the network using only training molecules and then evaluate on held-out paths.
  • A scaffold split or time-based split would stress-test whether evolutionary paths capture general structure-property trends or merely connect near-duplicate molecules that random splits place on both sides of the boundary.
  • The mutation-type interpretability suggests an untapped use: MEvoN could drive counterfactual search, proposing which single-atom or ring mutation most lowers a target property, rather than only improving regression accuracy.
  • The same evolutionary-path idea could apply to generative modeling, where an ancestral path is a natural sequence of edit operations for molecule optimization, not merely an auxiliary prediction feature.
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 / 5 minor

Summary. The paper proposes the Molecular Evolutionary Network (MEvoN), a similarity-based graph over molecules grouped by atom count, from which evolutionary paths are extracted. The proposed MEvoN-MPP method augments standard molecular encoders (GCN, GIN, SchNet, ComENet) with an EvoP module that encodes path features and an EvoL module that encodes property labels along the path, then predicts both the target property and the property difference between the last two molecules on the path. The authors report an average 32.3% MAE improvement over four base encoders on QM9 and competitive results on QM7, and they present ablations, SHAP analyses, and a path-number hyperparameter study.

Significance. If the reported gains were obtained under a leakage-free protocol, a plug-in that improves four diverse encoders by 32.3% average MAE on standard quantum-property benchmarks would be a meaningful contribution, and the public code is a strength. However, as described, the evaluation pipeline feeds known target-property labels of ancestor molecules into the EvoL module, and the MEvoN is constructed over the full dataset. The reported improvements therefore cannot currently be attributed to evolutionary-path representation; they may be due to label interpolation along the similarity graph. The ablations in Table 4 further show that the path-only EvoP module degrades performance, so the central claim is not supported by the evidence as presented.

major comments (3)
  1. [Section 3, Eqs. (14)-(15); Section 4.1] The evaluation is compromised by target-label information entering the prediction pipeline. In Section 3, the EvoL module receives the label sequence Y_i along each evolutionary path, and only the final label y_{l-1} is masked; y_pred2 is a function of H_EvoL and is trained against y_{l-2} - y_{l-1}. Section 4.1 states that each MEvoN is constructed from the full molecular dataset, and the 8:1:1 random split is applied to the dataset after network construction. The paper does not state that paths are restricted to training molecules or that ancestor labels are available only for training molecules. Consequently, for a test molecule, its evolutionary path can contain other test or validation molecules whose property labels are known inputs at test time. This turns the regression into label interpolation along the similarity graph rather than representation enhancement. The authors must rerun the experiments with a split-restricted protocol: construct the MEvoN using only training molecules, use only training labels in EvoL, and report test performance under that protocol. Without this, Tables 2-4 cannot be interpreted as evidence for the paper's central claim.
  2. [Table 4, Section 4.3] The ablation study directly contradicts the claim that evolutionary-path features enhance molecular representation. Adding the EvoP module alone worsens the GAP MAE from 0.2916 to 0.9809 on the Mol-branch and to 0.5187 on the Evo-branch, while adding EvoL alone improves the Mol-branch to 0.4051 and the combined system reaches 0.2270. Thus, the observed gain relative to the MolE baseline is driven by the label-sequence branch, not by the evolutionary-path representation. The manuscript should either reframe the contribution as a label-sequence augmentation method or provide a variant that obtains improvements from path features alone (e.g., without any property labels as inputs).
  3. [Section 4.1] The exclusion of 555 QM9 molecules from the MEvoN, including 446 fluorine-containing molecules and some N-containing cyclic structures, is a potential source of benchmark bias. These molecules are not reported in the evaluation, and the paper does not state whether they are excluded from both the training and test splits or only from the network construction. If the test set is restricted to molecules that happen to have high-similarity evolutionary neighbors, the reported MAE improvements may not generalize to the full QM9 distribution. The authors should report performance on the excluded molecules and on the full test set, or justify the exclusion as part of the method rather than a post hoc filter.
minor comments (5)
  1. [Section 3, first paragraph] The text introduces the module as 'label-aware (EvoA)' but the rest of the paper uses 'EvoL'; the notation should be made consistent.
  2. [Section 4.1] The sentence 'Figure 1 shows the number of molecules in the datasets...' appears to refer to Table 1, not Figure 1; the cross-reference should be corrected.
  3. [Table 4] The table caption and column headers do not define 'Mol-branch' and 'Evo-branch'; these terms should be linked explicitly to the MolE and Evo branches in Section 3, Eqs. (13)-(15).
  4. [Section 4.2, Table 2] The reported 'average of 32.3%' is an average of the four model-specific mean improvements; the paper should state this aggregation explicitly and provide per-seed standard deviations, since only seed 42 is reported for QM9.
  5. [Section 4.4, Figure 6] The path-number experiment reports no error bars or repeated-seed statistics, and it is unclear whether K was selected on the validation set; a single deterministic run is insufficient to support the stability claim for K between 3 and 5.

Circularity Check

1 steps flagged · score 6.0 of 10

Reported 32.3% gain may be label leakage: EvoL consumes ground-truth labels of full-dataset evolutionary paths, and only the final label is masked.

  1. fitted input called prediction [Section 3, Eqs. 13-15; Section 4 (data split); Section 4.1 (network construction over full datasets)]
    "The EvoL input is the molecular properties in P i k, denoted as Yi ∈ R L. In this case, yi l−1 and yi l−2 represent the labels of the last and the second-to-last valid molecules of Yi, respectively. The label of each path's last valid molecule, yi l−1, is masked to prevent data leakage."

    MEvoN is built over the entire dataset ('we constructed three evolutionary networks based on the QM7, QM8, and QM9 datasets') before the data are 'randomly split into training, validation, and test sets with a ratio of 8:1:1'. With a random split, an ancestor molecule on a test molecule's evolutionary path can itself be in validation or test, and its property label is then available as an input to EvoL. Eq. 14 puts HEvoL = EvoL(Yi) directly into y_pred2, so the model is handed known values of the target property for structurally related molecules. The paper masks only the last path label, not all non-training path labels.

full rationale

The construction of the MEvoN graph itself is self-contained: grouping by atom count, similarity thresholds, and path backtracking do not reference property labels. The circularity is in the evaluation protocol. The EvoL module is fed property labels along each evolutionary path, and paths are extracted from a network constructed over the full dataset before the random 8:1:1 split. Because the split is random, the path of a test molecule can contain validation or test molecules whose labels are then used as model inputs. Eq. 14 places HEvoL directly in the prediction, so the reported average 32.3% improvement is not attributable solely to evolutionary path structure; it may be label interpolation. The ablation supports this concern: path-only EvoP degrades MAE from 0.2916 to 0.9809, while the improvement appears only when the label-consuming EvoL is included. No statement in the paper restricts path labels to the training split. This is an evaluation-level circularity rather than a mathematical self-definition, so the score is 6 rather than higher.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The central claim rests on two categories of unpaid inputs. The first is the network construction: an atom-count hierarchy plus similarity thresholds θ=0.3 defines which molecules are evolutionarily related, and 555 QM9 molecules are carved out when they do not fit. The second is the label-aware module, which feeds known property values of path molecules into the predictor; if those labels include molecules from the test split, the reported improvements are inflated. No new physical entities are introduced.

free parameters (3)
  • Similarity thresholds θ1 and θ2 = 0.3 (both)
    Algorithm 1 uses two thresholds to decide evolutionary edges; Section 4.1 states thresholds were set to 0.3 with no sensitivity analysis or ablation.
  • Path number K = not reported (tested 1, 3, 5, 7, 9)
    The model selects up to K evolutionary paths per molecule; Figure 6 suggests K=3-5 is stable, but the final K used in Tables 2-3 is not stated.
  • Loss weights α and β = 1, 1
    Eq. 15 combines two MSE terms with weights α and β; Section 4 states the defaults are both 1.
assumptions (5)
  • domain assumption Molecules with fewer atoms are ancestral to larger molecules.
    Section 2.1 groups molecules by atom count and only connects groups i<j, imposing a directional ancestry that is not chemically established.
  • domain assumption Similarity measured by Tanimoto, edit distance, and WL kernel captures evolutionary relatedness.
    Section 2.2 uses these metrics to define evolutionary edges; there is no validation against known chemical or biological evolutionary relationships.
  • ad hoc to paper The property labels of ancestor molecules are available at inference time.
    The EvoL module encodes the label sequence Y_i of each path (Section 3, Eq. 14-15); the paper does not restrict labels to the training split.
  • ad hoc to paper Excluding low-similarity molecules does not bias the benchmark.
    Section 4.1 excludes 555 QM9 molecules, including 446 F-containing molecules, because they fit the network poorly; if baselines run on the full set, the comparison is inconsistent.
  • standard math A Transformer can usefully summarize path and label sequences.
    The method relies on standard Transformer behavior; this is not independently verified beyond the reported experiments.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Can Molecular Evolution Mechanism Enhance Molecular Representation?." pith.science (2026). https://pith.science/paper/FXVOU43F

@misc{pith2026250115799,
  author       = {Pith},
  title        = {Pith review of: Can Molecular Evolution Mechanism Enhance Molecular Representation?},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FXVOU43F}},
  note         = {Machine review of arXiv:2501.15799}
}
read the original abstract

Molecular evolution is the process of simulating the natural evolution of molecules in chemical space to explore potential molecular structures and properties. The relationships between similar molecules are often described through transformations such as adding, deleting, and modifying atoms and chemical bonds, reflecting specific evolutionary paths. Existing molecular representation methods mainly focus on mining data, such as atomic-level structures and chemical bonds directly from the molecules, often overlooking their evolutionary history. Consequently, we aim to explore the possibility of enhancing molecular representations by simulating the evolutionary process. We extract and analyze the changes in the evolutionary pathway and explore combining it with existing molecular representations. Therefore, this paper proposes the molecular evolutionary network (MEvoN) for molecular representations. First, we construct the MEvoN using molecules with a small number of atoms and generate evolutionary paths utilizing similarity calculations. Then, by modeling the atomic-level changes, MEvoN reveals their impact on molecular properties. Experimental results show that the MEvoN-based molecular property prediction method significantly improves the performance of traditional end-to-end algorithms on several molecular datasets. The code is available at https://anonymous.4open.science/r/MEvoN-7416/.

Figures

Figures reproduced from arXiv: 2501.15799 by the authors.

Figure 1
Figure 1. Molecular evolutionary network (MEvoN) illustrating the [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Evolutionary paths and molecular property changes for [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. The MEvoN method’s construction process. The steps are: 1) group molecules by atom count; 2) calculate inter-group similarity; [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Overview of MEvoN-MPP, which employs the MEvoN [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]
Figure 5
Figure 5. Figure 5: Importance of different mutation types in molecular evo [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: Hyperparameter experiment exploring the impact of dif [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

34 extracted references · 31 canonical work pages

  1. [1]

    Molecular modeling in drug discovery

    [Adelusi et al., 2022] Temitope Isaac Adelusi, Abdul- Quddus Kehinde Oyedele, Ibrahim Damilare Boyenle, Abdeen Tunde Ogunlana, Rofiat Oluwabusola Adeyemi, Chiamaka Divine Ukachi, Mukhtar Oluwaseun Idris, Olamide Tosin Olaoba, Ibrahim Olaide Adedotun, Oladipo Elijah Kolawole, et al. Molecular modeling in drug discovery. Informatics in Medicine Unlocked , 2...

    work page 2022

  2. [6]

    Reinforced genetic algorithm for structure-based drug design

    [Fu et al., 2022] Tianfan Fu, Wenhao Gao, Connor Coley, and Jimeng Sun. Reinforced genetic algorithm for structure-based drug design. Advances in Neural Infor- mation Processing Systems, 35:12325–12338,

    work page 2022

  3. [7]

    Molecular evo- lution meets the genomics revolution

    [H and H., 2003] Wolfe K H and Li W H. Molecular evo- lution meets the genomics revolution. Nature genetics, 33(3):255–265,

    work page 2003

  4. [13]

    [Liu et al., 2019] Pengfei Liu, Hongjian Li, Shuai Li, and Kwong Sak Leung

    Main Track. [Liu et al., 2019] Pengfei Liu, Hongjian Li, Shuai Li, and Kwong Sak Leung. Improving prediction of phenotypic drug response on cancer cell lines using deep convolu- tional network. BMC Bioinformatics, 20(1):1–14,

    work page 2019

  5. [14]

    Lundberg and Su-In Lee

    [Lundberg and Lee, 2017] Scott M. Lundberg and Su-In Lee. A unified approach to interpreting model predic- tions. In Proceedings of the 31st International Conference on Neural Information Processing Systems, NIPS’17, page 4768–4777. Curran Associates Inc.,

    work page 2017

  6. [15]

    Predicting compound activity from phe- notypic profiles and chemical structures

    [Moshkov et al., 2023] Nikita Moshkov, Tim Becker, Kevin Yang, Peter Horvath, Vlado Dancik, Bridget K Wagner, Paul A Clemons, Shantanu Singh, Anne E Carpenter, and Juan C Caicedo. Predicting compound activity from phe- notypic profiles and chemical structures. Nature commu- nications, 14(1):1967,

    work page 2023

  7. [16]

    Accurate ring strain energy calculations on saturated three-membered heterocycles with one group 13–16 element

    [Planells and Ferao, 2020] Alicia Rey Planells and Ar- turo Espinosa Ferao. Accurate ring strain energy calculations on saturated three-membered heterocycles with one group 13–16 element. Inorganic Chemistry , 59(16):11503–11513,

    work page 2020

  8. [20]

    Enumeration of 166 billion organic small molecules in the chemical universe database gdb-17

    [Ruddigkeit et al., 2012] Lars Ruddigkeit, Ruud Van Deursen, Lorenz C Blum, and Jean-Louis Reymond. Enumeration of 166 billion organic small molecules in the chemical universe database gdb-17. Journal of chemical information and modeling, 52(11):2864–2875,

    work page 2012

Show all 34 references
  1. [22]

    E(n) equivariant graph neural networks

    [Satorras et al., 2021] V´ıctor Garcia Satorras, Emiel Hooge- boom, and Max Welling. E(n) equivariant graph neural networks. In Marina Meila and Tong Zhang, editors, Pro- ceedings of the 38th International Conference on Machine Learning, volume 139 of Proceedings of Machine Le...

    2021

  2. [24]

    Schnet–a deep learning architecture for molecules and ma- terials

    [Sch¨utt et al., 2018] Kristof T Sch ¨utt, Huziel E Sauceda, P- J Kindermans, Alexandre Tkatchenko, and K-R M ¨uller. Schnet–a deep learning architecture for molecules and ma- terials. The Journal of Chemical Physics, 148(24),

    2018

  3. [26]

    Weisfeiler-lehman graph kernels

    [Shervashidze et al., 2011] Nino Shervashidze, Pascal Schweitzer, Erik Jan Van Leeuwen, Kurt Mehlhorn, and Karsten M Borgwardt. Weisfeiler-lehman graph kernels. Journal of Machine Learning Research, 12(9),

    2011

  4. [28]

    Chemical space travel

    [van Deursen and Reymond, 2007] Ruud van Deursen and Jean-Louis Reymond. Chemical space travel. ChemMed- Chem: Chemistry Enabling Drug Discovery , 2(5):636– 640,

    2007

  5. [30]

    Leveraging language model for advanced multiproperty molecular op- timization via prompt engineering

    [Wu et al., 2024] Zhenxing Wu, Odin Zhang, Xiaorui Wang, Li Fu, Huifeng Zhao, Jike Wang, Hongyan Du, Dejun Jiang, Yafeng Deng, Dongsheng Cao, et al. Leveraging language model for advanced multiproperty molecular op- timization via prompt engineering. Nature Machine Intel- lige...

    2024

  6. [31]

    Pushing the boundaries of molecular representation for drug discovery with the graph attention mechanism

    [Xiong et al., 2019] Zhaoping Xiong, Dingyan Wang, Xi- aohong Liu, Feisheng Zhong, Xiaozhe Wan, Xutong Li, Zhaojun Li, Xiaomin Luo, Kaixian Chen, Hualiang Jiang, et al. Pushing the boundaries of molecular representation for drug discovery with the graph attention mechanism. Jo...

    2019

  7. [32]

    How powerful are graph neural net- works? In International Conference on Learning Rep- resentations,

    [Xu et al., 2019] Keyulu Xu, Weihua Hu, Jure Leskovec, and Stefanie Jegelka. How powerful are graph neural net- works? In International Conference on Learning Rep- resentations,

    2019

  8. [33]

    Analyzing learned molecular representa- tions for property prediction

    [Yang et al., 2019] Kevin Yang, Kyle Swanson, Wengong Jin, Connor Coley, Philipp Eiden, Hua Gao, Angel Guzman-Perez, Timothy Hopper, Brian Kelley, Miriam Mathea, et al. Analyzing learned molecular representa- tions for property prediction. Journal of chemical infor- mation and...

    2019

  9. [34]

    Uni-mol: A universal 3d molecular representation learning framework

    [Zhou et al., 2023] Gengmo Zhou, Zhifeng Gao, Qiankun Ding, Hang Zheng, Hongteng Xu, Zhewei Wei, Linfeng Zhang, and Guolin Ke. Uni-mol: A universal 3d molecular representation learning framework. In The Eleventh Inter- national Conference on Learning Representations, 2023

    2023

  10. [1999]

    Quantum chemistry structures and properties of 134 kilo molecules

    [Ramakrishnan et al., 2014] Raghunathan Ramakrishnan, Pavlo O Dral, Matthias Rupp, and O Anatole V on Lilien- feld. Quantum chemistry structures and properties of 134 kilo molecules. Scientific data, 1(1):1–7,

    2014

  11. [2001]

    Geometry-enhanced molecu- lar representation learning for property prediction

    [Fang et al., 2022] Xiaomin Fang, Lihang Liu, Jieqiong Lei, Donglong He, Shanzhuo Zhang, Jingbo Zhou, Fan Wang, Hua Wu, and Haifeng Wang. Geometry-enhanced molecu- lar representation learning for property prediction. Nature Machine Intelligence, 4(2):127–134,

    2022

  12. [2003]

    Sofroniew, Deniz Oktay, Zeming Lin, Robert Verkuil, Vincent Q

    [Hayes et al., 2025] Thomas Hayes, Roshan Rao, Halil Akin, Nicholas J. Sofroniew, Deniz Oktay, Zeming Lin, Robert Verkuil, Vincent Q. Tran, Jonathan Deaton, Mar- ius Wiggert, Rohil Badkundri, Irhum Shafkat, Jun Gong, Alexander Derry, Raul S. Molina, Neil Thomas, Yousuf A. Khan...

    2025

  13. [2006]

    Fragment-masked molecular optimization

    [Li et al., 2024a] Kun Li, Xiantao Cai, Jia Wu, Bo Du, and Wenbin Hu. Fragment-masked molecular optimization. arXiv preprint arXiv:2408.09106,

    arXiv

  14. [2007]

    Phenotypic drug discovery: recent successes, lessons learned and new directions

    [Vincent et al., 2022] Fabien Vincent, Arsenio Nueda, Jonathan Lee, Monica Schenone, Marco Prunotto, and Mark Mercola. Phenotypic drug discovery: recent successes, lessons learned and new directions. Nature Reviews Drug Discovery, 21(12):899–914,

    2022

  15. [2011]

    Design-bench: Bench- marks for data-driven offline model-based optimization

    [Trabucco et al., 2022] Brandon Trabucco, Xinyang Geng, Aviral Kumar, and Sergey Levine. Design-bench: Bench- marks for data-driven offline model-based optimization. In International Conference on Machine Learning , pages 21658–21676. PMLR,

    2022

  16. [2012]

    [Rupp et al., 2012] M. Rupp, A. Tkatchenko, K.-R. M ¨uller, and O. A. von Lilienfeld. Fast and accurate modeling of molecular atomization energies with machine learning. Physical Review Letters, 108:058301,

    2012

  17. [2014]

    Self-supervised graph transformer on large-scale molecular data

    [Rong et al., 2020] Yu Rong, Yatao Bian, Tingyang Xu, Weiyang Xie, Ying Wei, Wenbing Huang, and Junzhou Huang. Self-supervised graph transformer on large-scale molecular data. Advances in neural information process- ing systems, 33:12559–12571,

    2020

  18. [2017]

    The molecule evolua- tor

    [Lameijer et al., 2006] Eric-Wubbo Lameijer, Joost N Kok, Thomas B¨ack, and Ad P IJzerman. The molecule evolua- tor. an interactive evolutionary algorithm for the design of drug-like molecules. Journal of chemical information and modeling, 46(2):545–552,

    2006

  19. [2018]

    Diffuse, sample, project: plug-and-play controllable graph generation

    [Sharma et al., 2025] Kartik Sharma, Srijan Kumar, and Rakshit S Trivedi. Diffuse, sample, project: plug-and-play controllable graph generation. In Proceedings of the 41st International Conference on Machine Learning, ICML’24. JMLR.org,

    2025

  20. [2019]

    Interchain interactions in or- ganic π-conjugated materials: impact on electronic struc- ture, optical response, and charge transport.Advanced ma- terials, 13(14):1053–1067,

    [Cornil et al., 2001] J´erˆome Cornil, David Beljonne, J-P Calbert, and J-L Br ´edas. Interchain interactions in or- ganic π-conjugated materials: impact on electronic struc- ture, optical response, and charge transport.Advanced ma- terials, 13(14):1053–1067,

    2001

  21. [2020]

    Electronic Processes in Organic Crystals and Polymers

    [Pope and Swenberg, 1999] Martin Pope and Charlese E Swenberg. Electronic Processes in Organic Crystals and Polymers. Oxford University Press, 12

    1999

  22. [2021]

    Schnet: A continuous-filter convolutional neural network for model- ing quantum interactions

    [Sch¨utt et al., 2017] Kristof Sch ¨utt, Pieter-Jan Kindermans, Huziel Enoc Sauceda Felix, Stefan Chmiela, Alexan- dre Tkatchenko, and Klaus-Robert M ¨uller. Schnet: A continuous-filter convolutional neural network for model- ing quantum interactions. Advances in neural inform...

    2017

  23. [2022]

    Regression transformer enables concurrent sequence re- gression and generation for molecular language modelling

    [Born and Manica, 2023] Jannis Born and Matteo Manica. Regression transformer enables concurrent sequence re- gression and generation for molecular language modelling. Nature Machine Intelligence, 5(4):432–444,

    2023

  24. [2023]

    Jac- card/tanimoto similarity test and estimation methods for biological presence-absence data

    [Chung et al., 2019] Neo Christopher Chung, Bła ˙Zej Mi- asojedow, Michał Startek, and Anna Gambin. Jac- card/tanimoto similarity test and estimation methods for biological presence-absence data. BMC bioinformatics , 20(Suppl 15):644,

    2019

  25. [2024]

    Regressor-free molecule gener- ation to support drug response prediction

    [Li et al., 2024b] Kun Li, Xiuwen Gong, Shirui Pan, Jia Wu, Bo Du, and Wenbin Hu. Regressor-free molecule gener- ation to support drug response prediction. arXiv preprint arXiv:2405.14536,

    arXiv

  26. [2025]

    Strategies for pre-training graph neural networks

    [Hu et al., 2020] W Hu, B Liu, J Gomes, M Zitnik, P Liang, V Pande, and J Leskovec. Strategies for pre-training graph neural networks. In International Conference on Learning Representations (ICLR),

    2020

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.