Pith. sign in

REVIEW 2 major objections 5 minor 54 references

Multi-Granular Rationale-Guided Molecular LLM for Property Prediction

T0 review · 2 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read This paper claims that a molecular LLM's property predictions improve measurably, and become more interpretable, when the prompt includes a ranked, direction-tagged list of GNN-derived substructure attributions, and that the model reads…

desk verdict Genuinely new mechanism and convincing diagnostics, but the 'best generalist' headline is undercut by an unfair zero-shot vs fine-tuned comparison. read the letter →

arxiv 2608.10480 v1 pith:OZSVKF4K submitted 2026-08-11 cs.AI cs.LG

Junwoo Park , Minyoung Shin , Cheol Soon Lee , Sujee Lee This is my paper · ORCID
classification cs.AIcs.LG
keywords molecularpropertypredictionlargelanguagemodelsgraphneuralnetworkssubstructuremaskingexplainableAImulti-granularrationaleMoleculeNetrationale-guidedLLM
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

MR-MoL tests a simple proposal: give a molecular LLM the same evidence a chemist would use — the substructures that push a property up or down — and the model's property predictions improve while becoming legible. The paper claims that a fine-tuned GNN's masking attributions, serialized as a ranked, direction-tagged, multi-granular rationale, let a generalist LLM beat all generalist baselines on most of eight MoleculeNet tasks and narrow the gap to per-task specialists. It further claims the LLM genuinely reads this rationale: flipping the direction tag reverses predictions, removing the top-ranked item shifts the output several times more than removing a lower-ranked one, and replacing a substructure with a random one hurts performance. If these claims hold, the route from black-box GNN explanations to LLM evidence is a practical way to make molecular property prediction both stronger and more interpretable.

What carries the argument

The load-bearing object is the multi-granular rationale: a ranked, direction-tagged textual list of the top five substructures from three complementary views — Murcko scaffolds with side chains, BRICS fragments, and chemically named functional groups — each scored by the masking attribution $a_{t,j} = f_t(G) - f_t(G \setminus u_j)$ of a per-task fine-tuned GNN. This rationale is concatenated into the prompt beside the SMILES string and projected graph tokens, so the LLM conditions directly on per-substructure evidence. Everything that distinguishes MR-MoL from a plain graph-aligned LLM passes through this channel.

What would settle it

Train the source predictor on randomly shuffled labels, so the masking attributions are pure noise, and feed the resulting 'rationale' to MR-MoL; if predictions do not fall back to (or below) the no-rationale baseline, the model is keying on the rationale's form rather than its chemical content — the diagnostics in the paper only perturb content, not source-predictor quality.

Watch

Extended reading notes

Core claim

The central discovery is that GNN-derived substructure attributions, when serialized as text and placed in the prompt, behave as real evidence for an instruction-tuned molecular LLM, not as inert filler. MR-MoL computes, for each of three decomposition views, the masking attribution $a_{t,j} = f_t(G) - f_t(G \setminus u_j)$ from a task-fine-tuned Mole-BERT source predictor, keeps the five largest-magnitude items, and tags each as pushing the predicted property toward higher or lower. The paper reports that this rationale is responsible for the model's best-overall-generalist results on eight MoleculeNet tasks, and that five diagnostics show the model's predictions move with the direction tag, the rank, and the substructure identity of each rationale item. On the chemical side, the attributions reproduce known structure-property relationships, such as polar functional groups raising aqueous solubility and lowering lipophilicity, and the rationale can flip an individual wrong prediction to the correct one. The method is presented as the first to expose GNN-derived attributions to an LLM as input evidence for property prediction.

Load-bearing premise

The whole rationale channel inherits its content from a single fine-tuned GNN's masking attributions; if those attributions are noisy, biased by the training split, or wrong for a particular molecule, the rationale becomes misinformation that can hurt predictions, and the paper's limitations section concedes exactly this risk.

Editorial extensions

If this is right

  • A single generalist molecular LLM can approach per-task specialist performance when its prompt carries explicit substructure evidence, without per-task fine-tuning of the LLM itself.
  • The rationale's direction, rank, and substructure are all causally read by the model; any one of these interventions moves the prediction, so the channel is informative rather than decorative.
  • Removing the rationale degrades performance on seven of eight tasks, with the largest drops on ClinTox and BACE, so the substructure evidence carries information the graph and SMILES channels do not.
  • The attributions reproduce known chemistry — polar functional groups raise predicted solubility and lower predicted lipophilicity — which suggests the evidence channel is aligned with domain knowledge.
  • On individual molecules, the rationale can correct a wrong prediction (e.g., nalidixic acid crossing the blood-brain barrier), and the correction depends on the top-attributed item.

Reading between the lines

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

  • Because the rationale is task-conditioned and trained, the same channel could plausibly be extended to tasks beyond classification and regression — such as reaction outcome prediction — if a source predictor can produce per-fragment attributions for those settings.
  • The diagnostics suggest the model may be using the rationale as a form of soft chain-of-thought; a testable extension would be to prompt the model to verbalize why a given direction tag from the GNN should or should not apply, and to measure whether such self-verification further improves accuracy.
  • If attribution noise (rather than the LLM) is the bottleneck, ensembling multiple source predictors or using more stable attribution methods than raw masking differences could push MR-MoL past the specialist GNNs on regression, where the gap currently remains.
  • The fixed top-five cutoff across all three views may be suboptimal; a task-adaptive selection of granularity and number of items could improve both performance and sparsity of the rationale.
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

2 major / 5 minor

Summary. The paper proposes MR-MoL, a multi-modal molecular LLM that augments a Llama-3.1-8B backbone with a 2D graph embedding path and a multi-granular, ranked, direction-tagged rationale derived from a fine-tuned GNN's substructure masking attributions. The rationale draws on Murcko scaffolds, BRICS fragments, and functional groups, and is serialized as text alongside the SMILES string and graph tokens. Training proceeds in two stages: graph-language alignment (Stage 1) and multi-task rationale-guided instruction tuning on eight MoleculeNet tasks (Stage 2). The central claims are (i) MR-MoL achieves the best overall results among generalist models and narrows the gap to task-tuned specialists, and (ii) the LLM genuinely reads the rationale, supported by five diagnostics: direction flipping, rank-1 vs. random removal, substructure shuffling, chemical validity of attributions, and individual prediction corrections.

Significance. If the claims hold, the paper introduces a useful and interpretable mechanism: exposing GNN-derived per-substructure attributions to an LLM as auxiliary evidence for property prediction. The internal experimental design is strong: the ablations isolate the rationale and graph channels, and the diagnostics are thoughtful and convincing. In particular, direction flipping systematically drives classification below chance (Table 3), rank sensitivity shows a 1.6–6.2× larger effect for the top item (Table 4), and substructure shuffling degrades all four tested tasks (Table 5). The chemical-validity check (Table 6) provides an additional, interpretability-oriented validation. These experiments support the claim that the model uses the rationale's content. However, the external comparison against 'generalist' baselines is not fair as presented: MR-MoL receives Stage 2 supervision on the eight evaluation tasks, whereas the generalist baselines are evaluated zero-shot from released checkpoints.

major comments (2)
  1. [Implementation Details and Table 18 (also Appendix C.3)] The generalist comparison is not apples-to-apples. As stated in Implementation Details, 'generalist models are evaluated from their publicly available checkpoints,' while MR-MoL is trained in Stage 2 on the eight evaluation tasks. Table 18 makes the asymmetry explicit: each generalist baseline has seen a different subset of the tasks, and several tasks are completely absent from some models' training data (e.g., ChemDFM is marked absent on BBBP, Tox21, ESOL, and Lipo). The abstract's claim that MR-MoL 'achieves the best overall results among generalist models' is therefore a comparison of a task-fine-tuned model against zero-shot models, which does not support the stated conclusion. To make the claim meaningful, the authors should either evaluate MR-MoL in a zero-shot setting (without Stage 2 supervision on the eight tasks), fine-tune the generalist baselines under an identical Stage 2 protocol, or explicitly reframe the conclusion as 'best among models evaluated under our multi-task instruction tuning' rather than 'best among generalist models.'
  2. [Results and Analysis, 'Comparison with baselines' and 'Ablation' paragraphs] The attribution of the generalist gains is internally inconsistent with the ablation evidence. In the 'Comparison with baselines' paragraph, the authors write that the advantage over generalists is due to 'the structural evidence MR-MoL receives alongside the 1D and 2D molecule representations.' However, Table 2 shows that the w/oG,R variant, which removes both the graph and the rationale and retains only the instruction, the SMILES string, and the same Stage 2 multi-task instruction tuning, already surpasses the generalist baselines in Table 1 on most tasks (as the paper itself acknowledges in the Ablation paragraph). This demonstrates that the dominant contributor to the generalist gap is the strong LLM backbone plus task-level supervision, not the rationale or graph channels. The internal comparison of MR-MoL against w/oR (which isolates the rationale under the same training protocol) remains valid and is compelling, but the external generalist claim as stated is unsupported by the presented evidence. This is a load-bearing issue because it affects the abstract's primary quantitative claim.
minor comments (5)
  1. [Table 4 (Rank sensitivity)] The p-values in Table 4 are formatted abnormally: '5.6×2×10−27' and '4×10−22' include an extra '2' that appears spurious. These should be corrected to a standard scientific notation (e.g., 5.6×10−27). The 'ratio' column is clear but the p-values need fixing.
  2. [Direction flipping (Table 3) text] In the direction-sensitivity subsection, the statement 'the ROC-AUC falls below chance' would benefit from an explicit clarification that values below 0.5 indicate that the model is now systematically following the inverted direction tag, since readers may need to connect this to the sign-flipped MCC.
  3. [Appendix C.3] The sentence 'We therefore avoid drawing conclusions from any single cell, and rely on the overall comparison rather than task-by-task claims against these baselines' is a reasonable defense against per-cell cherry-picking, but the overall comparison is still confounded by the training-time asymmetry. This caveat should be moved or echoed in the main text, not only the Appendix, because it directly affects the interpretation of the headline result.
  4. [Table 1 caption] The caption states that the overall best across groups is underlined, but in the table as rendered no underlining is visible. Please ensure the formatting is correct in the final version.
  5. [Method, Ranked Multi-Granular Rationale] When defining the attribution a_{t,j} in Eq. (4), the exact meaning of masking G\u_j for each view (Murcko, BRICS, functional group) is left to the reader. A one-sentence example or a pointer to Appendix B for the precise masking operations would improve clarity.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the rationale is computed by an independent masked-GNN attribution without access to test labels, and the diagnostics independently establish that the LLM uses it.

full rationale

The paper's central derivation is not circular. The rationale is produced by a separately fine-tuned GNN via masking attribution in Eq. (4), a_{t,j} = f_t(G) - f_t(G \ u_j), which is computed for test molecules without using their labels; the LLM is then trained with the standard next-token objective in Eq. (6). The claim that the model reads the rationale rather than ignoring it is tested by interventions that are not encoded in the training objective: flipping direction tags (Table 3), removing rank-1 items (Table 4), and shuffling substructures (Table 5) all move predictions in the expected directions. The self-citation to the authors' earlier LLaMo work is taxonomical background in Related Work and is not load-bearing for any derivation. The acknowledged limitation that the rationale inherits source-predictor quality is an error-propagation caveat, not a circularity. The benchmarking concern that MR-MoL is task-fine-tuned while generalist baselines are evaluated zero-shot is a fairness/correctness issue outside the circularity definition, since it does not make any prediction equivalent to its input by construction. Overall the derivation is self-contained and the diagnostic evidence supports the stated claims.

Assumptions & free parameters 1 free parameters · 4 assumptions · 0 invented entities

The paper does not introduce new physical entities or fitted physical constants. Its load-bearing assumptions are the meaningfulness of GNN masking attributions and the coverage of the three decomposition views. The only hand-chosen numeric design choice that affects the rationale is the top-k=5 selection.

free parameters (1)
  • k (number of rationale items per molecule) = 5
    The paper keeps the top five substructures by absolute attribution. Appendix B.3 compares Top-1 vs Full and finds Full better on most tasks, but k=5 is a hand-chosen hyperparameter, not an optimized constant.
assumptions (4)
  • domain assumption Substructure masking attribution (a_{t,j} = f_t(G) - f_t(G \ u_j)) faithfully ranks substructures by their contribution to the GNN prediction.
    The rationale is built entirely from these attributions (Eq. 4). If they are uninformative, the method reduces to adding noise to the prompt. The paper cites Wu et al. (2023) for this approach but does not prove its validity for each task.
  • domain assumption The three decomposition views (Murcko, BRICS, functional groups) jointly cover the substructures that drive the target properties.
    Substructures outside these views, such as macrocycles or stereochemistry, are excluded, as the paper notes in Limitations. If the key substructure is not in any view, the rationale cannot mention it.
  • domain assumption Next-token prediction training with LoRA can learn to condition on concatenated molecular embeddings and rationale text without catastrophic forgetting.
    The two-stage training assumes the frozen LLM backbone can integrate these modalities; the paper provides no formal guarantee, only empirical results.
  • domain assumption Scaffold split yields test molecules that are structurally distinct enough to measure generalization.
    This is the standard MoleculeNet protocol, but it is an assumption about the benchmark's difficulty and independence.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Multi-Granular Rationale-Guided Molecular LLM for Property Prediction." pith.science (2026). https://pith.science/paper/OZSVKF4K

@misc{pith2026260810480,
  author       = {Pith},
  title        = {Pith review of: Multi-Granular Rationale-Guided Molecular LLM for Property Prediction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OZSVKF4K}},
  note         = {Machine review of arXiv:2608.10480}
}
read the original abstract

Large language models (LLMs) are widely applied across chemical tasks, such as molecular property prediction, which underpins drug discovery. Molecular LLMs represent a molecule through several modalities, notably a 1D SMILES sequence or a 2D molecular graph. Both encode molecular information implicitly, so the contribution of individual substructures remains opaque. Retrieval and augmentation methods add context, but from external sources. However, the cues chemists reason over are the internal substructures that drive a property up or down. We propose MR-MoL, a multi-granular rationale-guided molecular LLM that supplies this evidence directly. A fine-tuned GNN scores each substructure through masking, and the most influential ones are serialized as a ranked, direction-tagged rationale that the LLM reads alongside the SMILES sequence and molecular graph. The rationale spans three levels of granularity: Murcko scaffolds with their side chains, BRICS fragments, and functional groups. This is, to our knowledge, the first method to expose GNN-derived attributions to an LLM as evidence for property prediction. On eight MoleculeNet tasks, MR-MoL achieves the best overall results among generalist models and narrows the gap to specialist models tuned for each task. Five diagnostics further confirm that the model reads the rationale rather than merely benefiting from its presence. Its direction, rank, and substructure each shape the prediction, and its attributions reproduce known structure-property relationships.

Figures

Figures reproduced from arXiv: 2608.10480 by the authors.

Figure 1
Figure 1. Overview of MR-MoL. The graph embedding path encodes a 2D molecular graph through a graph encoder and the [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. A multi-granular rationale corrects a blood-brain [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Chat template skeleton for both stages. The [PITH_FULL_IMAGE:figures/full_fig_p013_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Distribution of substructure attributions [PITH_FULL_IMAGE:figures/full_fig_p014_4.png]
Figure 5
Figure 5. Figure 5: The first three items of a serialized rationale as fed to the LLM, for one molecule from the ESOL dataset with SMILES [PITH_FULL_IMAGE:figures/full_fig_p015_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

54 extracted references · 34 canonical work pages

  1. [1]

    Anthropic . 2026. Claude Opus 4.6 System Card. https://www.anthropic.com/claude-opus-4-6-system-card. Accessed: 2026-06-09

    work page 2026

  2. [2]

    W.; and Murcko, M

    Bemis, G. W.; and Murcko, M. A. 1996. The properties of known drugs. 1. Molecular frameworks. Journal of medicinal chemistry, 39(15): 2887--2893

    1996

  3. [3]

    Cao, H.; Liu, Z.; Lu, X.; Yao, Y.; and Li, Y. 2025. Instructmol: Multi-modal integration for building a versatile and reliable molecular assistant in drug discovery. In Proceedings of the 31st International Conference on Computational Linguistics, 354--379

    work page 2025

  4. [4]

    Chen, Y.; Yao, Q.; Zhang, J.; Cheng, J.; and Bian, Y. 2024. Hight: Hierarchical graph tokenization for graph-language alignment. arXiv:2406.14021

    arXiv 2024

  5. [5]

    Chithrananda, S.; Grand, G.; and Ramsundar, B. 2020. ChemBERTa: large-scale self-supervised pretraining for molecular property prediction. arXiv:2010.09885

    arXiv 2020

  6. [6]

    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

    work page 2008

  7. [7]

    Edwards, C.; Zhai, C.; and Ji, H. 2021. Text2mol: Cross-modal molecule retrieval with natural language queries. In Proceedings of the 2021 conference on empirical methods in natural language processing, 595--607

    work page 2021

  8. [8]

    Fang, Y.; Liang, X.; Zhang, N.; Liu, K.; Huang, R.; Chen, Z.; Fan, X.; and Chen, H. 2024. Mol-Instructions: A Large-Scale Biomolecular Instruction Dataset for Large Language Models. In The Twelfth International Conference on Learning Representations

    work page 2024

Show all 54 references
  1. [9]

    Fang, Y.; Zhang, Q.; Zhang, N.; Chen, Z.; Zhuang, X.; Shao, X.; Fan, X.; and Chen, H. 2023. Knowledge graph-enhanced molecular contrastive learning with functional prompt. Nature Machine Intelligence, 5(5): 542--553

    2023

  2. [10]

    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 International conference on machine learning, 1263--1272. Pmlr

    2017

  3. [11]

    Grattafiori, A.; Dubey, A.; Jauhri, A.; Pandey, A.; Kadian, A.; Al-Dahle, A.; Letman, A.; Mathur, A.; Schelten, A.; Vaughan, A.; et al. 2024. The llama 3 herd of models. arXiv:2407.21783

    2024 arXiv

  4. [12]

    Hastings, J.; De Matos, P.; Dekker, A.; Ennis, M.; Harsha, B.; Kale, N.; Muthukrishnan, V.; Owen, G.; Turner, S.; Williams, M.; et al. 2012. The ChEBI reference database and ontology for biologically relevant chemistry: enhancements for 2013. Nucleic acids research, 41(D1): D456--D463

    2012

  5. [13]

    J.; Shen, Y.; Wallis, P.; Allen-Zhu, Z.; Li, Y.; Wang, S.; Wang, L.; and Chen, W

    Hu, E. J.; Shen, Y.; Wallis, P.; Allen-Zhu, Z.; Li, Y.; Wang, S.; Wang, L.; and Chen, W. 2022. LoRA: Low-Rank Adaptation of Large Language Models. In International Conference on Learning Representations

    2022

  6. [14]

    Hu*, W.; Liu*, B.; Gomes, J.; Zitnik, M.; Liang, P.; Pande, V.; and Leskovec, J. 2020. Strategies for Pre-training Graph Neural Networks. In International Conference on Learning Representations

    2020

  7. [15]

    Jain, N.; and Yalkowsky, S. H. 2001. Estimation of the aqueous solubility I: Application to organic nonelectrolytes. Journal of Pharmaceutical Sciences, 90(2): 234--252

    2001

  8. [16]

    Jim \'e nez-Luna, J.; Grisoni, F.; and Schneider, G. 2020. Drug discovery with explainable artificial intelligence. Nature Machine Intelligence, 2(10): 573--584

    2020

  9. [17]

    Jin, C.; Guo, S.; Zhou, S.; and Guan, J. 2025. Effective and explainable molecular property prediction by chain-of-thought enabled large language models and multi-modal molecular information fusion. Journal of Chemical Information and Modeling, 65(11): 5438--5455

    2025

  10. [18]

    Kim, D.; Lee, W.; and Hwang, S. J. 2026. Mol-llama: Towards general understanding of molecules in large molecular language model. Advances in Neural Information Processing Systems, 38: 26921--26960

    2026

  11. [19]

    M.; Franklin, M.; Oler, E.; Wilson, A.; Pon, A.; Cox, J.; Chin, N

    Knox, C.; Wilson, M.; Klinger, C. M.; Franklin, M.; Oler, E.; Wilson, A.; Pon, A.; Cox, J.; Chin, N. E.; Strawbridge, S. A.; et al. 2024. DrugBank 6.0: the DrugBank knowledgebase for 2024. Nucleic acids research, 52(D1): D1265--D1275

    2024

  12. [20]

    J.; Tran, V.; Vijayendran, K

    Lassalas, P.; Gay, B.; Lasfargeas, C.; James, M. J.; Tran, V.; Vijayendran, K. G.; Brunden, K. R.; Kozlowski, M. C.; Thomas, C. J.; Smith, I., Amos B; Huryn, D. M.; and Ballatore, C. 2016. Structure property relationships of carboxylic acid isosteres. Journal of Medicinal Chem...

    2016

  13. [21]

    Lee, N.; De Brouwer, E.; Hajiramezanali, E.; Biancalani, T.; Park, C.; and Scalia, G. 2026. Rag-enhanced collaborative llm agents for drug discovery. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 40, 561--569

    2026

  14. [22]

    Li, J.; Li, D.; Savarese, S.; and Hoi, S. 2023. Blip-2: Bootstrapping language-image pre-training with frozen image encoders and large language models. In International conference on machine learning, 19730--19742. PMLR

    2023

  15. [23]

    Li, S.; Liu, Z.; Luo, Y.; Wang, X.; He, X.; Kawaguchi, K.; Chua, T.-S.; and Tian, Q. 2024. Towards 3d molecule-text interpretation in language models. In International Conference on Learning Representations, volume 2024, 17352--17371

    2024

  16. [24]

    Liu, P.; Ren, Y.; Tao, J.; and Ren, Z. 2024 a . Git-mol: A multi-modal large language model for molecular science with graph, image, and text. Computers in biology and medicine, 171: 108073

    2024

  17. [25]

    Liu, S.; Wang, H.; Liu, W.; Lasenby, J.; Guo, H.; and Tang, J. 2022. Pre-training Molecular Graph Representation with 3D Geometry. In International Conference on Learning Representations

    2022

  18. [26]

    Liu, Y.; Ding, S.; Zhou, S.; Fan, W.; and Tan, Q. 2024 b . MolecularGPT: Open Large Language Model (LLM) for Few-Shot Molecular Property Prediction. arXiv:2406.12950

    2024 arXiv

  19. [27]

    Liu, Z.; Li, S.; Luo, Y.; Fei, H.; Cao, Y.; Kawaguchi, K.; Wang, X.; and Chua, T.-S. 2023. Molca: Molecular graph-language modeling with cross-modal projector and uni-modal adapter. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, 15623--15638

    2023

  20. [28]

    Livne, M.; Miftahutdinov, Z.; Tutubalina, E.; Kuznetsov, M.; Polykovskiy, D.; Brundyn, A.; Jhunjhunwala, A.; Costa, A.; Aliper, A.; Aspuru-Guzik, A.; et al. 2024. nach0: multimodal natural and chemical languages foundation model. Chemical Science, 15(22): 8380--8389

    2024

  21. [29]

    Loshchilov, I.; and Hutter, F. 2019. Decoupled Weight Decay Regularization. In International Conference on Learning Representations

    2019

  22. [30]

    C.; Hom, R

    Maillard, M. C.; Hom, R. K.; Benson, T. E.; Moon, J. B.; Mamo, S.; Bienkowski, M.; Tomasselli, A. G.; Woods, D. D.; Prince, D. B.; Paddock, D. J.; et al. 2007. Design, synthesis, and crystal structure of hydroxyethyl secondary amine-based peptidomimetic inhibitors of human -se...

    2007

  23. [31]

    Ouyang, L.; Wu, J.; Jiang, X.; Almeida, D.; Wainwright, C.; Mishkin, P.; Zhang, C.; Agarwal, S.; Slama, K.; Ray, A.; et al. 2022. Training language models to follow instructions with human feedback. Advances in neural information processing systems, 35: 27730--27744

    2022

  24. [32]

    Park, J.; Bae, M.; Ko, D.; and Kim, H. J. 2024. Llamo: Large language model-based molecular graph assistant. Advances in Neural Information Processing Systems, 37: 131972--132000

    2024

  25. [33]

    Pei, Q.; Yan, R.; Gao, K.; Zhu, J.; and Wu, L. 2025. 3D-MolT5: Leveraging Discrete Structural Information for Molecule-Text Modeling. In International Conference on Learning Representations

    2025

  26. [34]

    Ramsundar, B.; Eastman, P.; Walters, P.; Pande, V.; Leswing, K.; and Wu, Z. 2019. Deep Learning for the Life Sciences. O'Reilly Media. https://www.amazon.com/Deep-Learning-Life-Sciences-Microscopy/dp/1492039837

    2019

  27. [35]

    RDKit . 2006. RDKit: Open-source cheminformatics. https://www.rdkit.org. Accessed: 2026-05-06

    2006

  28. [36]

    Ross, J.; Belgodere, B.; Chenthamarakshan, V.; Padhi, I.; Mroueh, Y.; and Das, P. 2022. Large-scale chemical language representations capture molecular structure and properties. Nature Machine Intelligence, 4(12): 1256--1264

    2022

  29. [37]

    C.; Hagenbuchner, M.; and Monfardini, G

    Scarselli, F.; Gori, M.; Tsoi, A. C.; Hagenbuchner, M.; and Monfardini, G. 2008. The graph neural network model. IEEE transactions on neural networks, 20(1): 61--80

    2008

  30. [38]

    u tt, K. T.; Sauceda, H. E.; Kindermans, P.-J.; Tkatchenko, A.; and M \

    Sch \"u tt, K. T.; Sauceda, H. E.; Kindermans, P.-J.; Tkatchenko, A.; and M \"u ller, K.-R. 2018. Schnet--a deep learning architecture for molecules and materials. The Journal of chemical physics, 148(24)

    2018

  31. [39]

    T.; Hou, X.; Verhoest, P

    Wager, T. T.; Hou, X.; Verhoest, P. R.; and Villalobos, A. 2010. Moving beyond rules: the development of a central nervous system multiparameter optimization (CNS MPO) approach to enable alignment of druglike properties. ACS Chemical Neuroscience, 1(6): 435--449

    2010

  32. [40]

    Wang, Y.; Wang, J.; Cao, Z.; and Barati Farimani, A. 2022. Molecular contrastive learning of representations via graph neural networks. Nature Machine Intelligence, 4(3): 279--287

    2022

  33. [41]

    W.; Lester, B.; Du, N.; Dai, A

    Wei, J.; Bosma, M.; Zhao, V.; Guu, K.; Yu, A. W.; Lester, B.; Du, N.; Dai, A. M.; and Le, Q. V. 2022. Finetuned Language Models are Zero-Shot Learners. In International Conference on Learning Representations

    2022

  34. [42]

    P.; Gandhi, H

    Wellawatte, G. P.; Gandhi, H. A.; Seshadri, A.; and White, A. D. 2023. A perspective on explanations of molecular prediction models. Journal of Chemical Theory and Computation, 19(8): 2149--2160

    2023

  35. [43]

    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

  36. [44]

    Wu, Z.; Wang, J.; Du, H.; Jiang, D.; Kang, Y.; Li, D.; Pan, P.; Deng, Y.; Cao, D.; Hsieh, C.-Y.; et al. 2023. Chemistry-intuitive explanation of graph neural networks for molecular property prediction with substructure masking. Nature communications, 14(1): 2585

    2023

  37. [45]

    Xia, J.; Zhao, C.; Hu, B.; Gao, Z.; Tan, C.; Liu, Y.; Li, S.; and Li, S. Z. 2023. Mole-bert: Rethinking pre-training graph neural networks for molecules. In The Eleventh International Conference on Learning Representations

    2023

  38. [46]

    Xian, Z.; Gu, J.; Li, L.; and Liang, S. 2025. Molrag: unlocking the power of large language models for molecular property prediction. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 15513--15531

    2025

  39. [47]

    Ying, Z.; Bourgeois, D.; You, J.; Zitnik, M.; and Leskovec, J. 2019. Gnnexplainer: Generating explanations for graph neural networks. Advances in neural information processing systems, 32

    2019

  40. [48]

    N.; Chen, Z.; Ning, X.; and Sun, H

    Yu, B.; Baker, F. N.; Chen, Z.; Ning, X.; and Sun, H. 2024. Lla SM ol: Advancing Large Language Models for Chemistry with a Large-Scale, Comprehensive, High-Quality Instruction Tuning Dataset. In First Conference on Language Modeling

    2024

  41. [49]

    Yuan, H.; Yu, H.; Wang, J.; Li, K.; and Ji, S. 2021. On explainability of graph neural networks via subgraph explorations. In International conference on machine learning, 12241--12252. PMLR

    2021

  42. [50]

    Zhang, Z.; Liu, Q.; Wang, H.; Lu, C.; and Lee, C.-K. 2021. Motif-based graph self-supervised learning for molecular property prediction. Advances in Neural Information Processing Systems, 34: 15870--15882

    2021

  43. [51]

    Zhao, H.; Liu, S.; Ma, C.; Xu, H.; Fu, J.; Deng, Z.; Kong, L.; and Liu, Q. 2023. GIMLET: A unified graph-text model for instruction-based molecule zero-shot learning. Advances in Neural Information Processing Systems, 36: 5850--5887

    2023

  44. [52]

    Zhao, J.; Zhou, Q.; Wang, T.; Fan, Y.; Yang, Q.; Jiao, L.; Liu, C.; Guo, Z.; Lu, Q.; Zhou, F.; et al. 2025. MolProphecy: Bridging medicinal chemists’ knowledge and molecular pre-trained models via a multi-modal framework. Journal of Advanced Research

    2025

  45. [53]

    Zhao, Z.; Ma, D.; Chen, L.; Sun, L.; Li, Z.; Xia, Y.; Xu, H.; Zhu, Z.; Zhu, S.; Fan, S.; et al. 2024. Chemdfm: A large language foundation model for chemistry. In Neurips 2024 Workshop Foundation Models for Science: Progress, Opportunities, and Challenges

    2024

  46. [54]

    Zhou, G.; Gao, Z.; Ding, Q.; Zheng, H.; Xu, H.; Wei, Z.; Zhang, L.; and Ke, G. 2023. Uni-Mol: A Universal 3D Molecular Representation Learning Framework. In The Eleventh International Conference on Learning Representations

    2023

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.