Pith. sign in

REVIEW 3 major objections 5 minor 66 references

MolecularCanvas: LLM-assisted Small-Molecule Drug Discovery via Structure-Guided Constraints

T0 review · 3 major / 5 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read MolecularCanvas claims that letting chemists annotate molecules directly on a canvas makes LLM-assisted drug optimization more effective, with expert-rated final molecules beating a familiar-tool baseline.

desk verdict The system design is thoughtful and the qualitative findings are plausible, but the paired user study is confounded by task, so the headline quantitative claims don't stand as reported. read the letter →

arxiv 2608.00393 v1 pith:HYHIEO3X submitted 2026-08-01 cs.HC

Haoyu Dong , Rui Sheng , Shuhao Zhang , Yushi Sun , Dingyang Wu , Hanxiang Chao , Olexandr Isayev , Huamin Qu
show 2 more authors
Yuyang Wu Yanna Lin
This is my paper
classification cs.HC
keywords largelanguagemodelsmolecularoptimizationsmall-moleculedrugdiscoverystructure-guidedconstraintshuman-AIinteractioninteractivesystemsin-contextpromptingcandidategeneration
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

MolecularCanvas is an interactive system for AI-assisted small-molecule optimization. The paper's claim is that letting chemists build a structured optimization context—high-level goals, on-molecule annotations (anchor, avoid, free-form edits), property target ranges, and reference-based preferences—makes LLM-generated candidates align better with expert intent than single free-form prompts. It also claims that attaching evidence from public drug databases to each candidate and integrating property calculators into one interface builds trust and removes the need to switch between separate tools. A 12-participant user study is offered as evidence: MolecularCanvas was rated significantly higher than the familiar-tool baseline on all 13 questionnaire items, scored higher on the System Usability Scale, and its final molecules were rated higher by two independent medicinal chemists (normalized 0.73 vs 0.58, p<0.001). The paper further argues this points to a broader design principle: in structure-intensive domains, the domain object itself should be the medium for instructing generative AI.

What carries the argument

The load-bearing mechanism is the structured optimization context plus a two-stage generation pipeline. The context is a formal specification that captures the current molecule in SMILES, a text goal, property constraints with target ranges, and atom-resolved structural annotations (anchor/avoid/free-form) taken directly from the Ketcher molecular editor, without vision-language parsing. The pipeline first asks an LLM to propose an edit plan as discrete operations, then runs those operations through RDKit to generate concrete molecules, checks chemical validity, removes duplicates and near-duplicates by Morgan fingerprint similarity (Tanimoto >0.85), scores candidates on property satisfactio

What would settle it

A matched controlled experiment would settle it: same LLM, same property calculators, same evidence database, and only the input modality varied (structured canvas annotations vs. plain text prompts), with chemists randomly assigned and unaware of the hypothesis. If expert-rated final-molecule quality, structural-constraint satisfaction, and SUS scores are not higher in the structured condition, the central claim fails. A cheaper log-based check: track whether anchor/avoid constraints are ever violated by top-ranked candidates in the current system; if violations are rare even in text-only bas

Watch

Extended reading notes

Core claim

At the center of the paper is a proposed fix for a workflow mismatch: chemists think in molecular structures, but generative AI tools mostly accept language. MolecularCanvas lets users select atoms and bonds on a 2D canvas and tag them as anchor (preserve), avoid (do not touch), or free-form instructions, combine this with desired property ranges and reference molecules, and feed the whole structured package to an LLM. The LLM is asked to produce an explicit edit plan of executable operations—replace atom X with fluorine, add a polar group at position Y—rather than a finished molecule, and RDKit executes and validates the plan. The paper's claim is that this structured, evidence-backed, inte

Load-bearing premise

The load-bearing premise is that the baseline—each participant using whatever familiar tools they chose—is a fair comparator, so the measured gains reflect MolecularCanvas's design rather than unequal familiarity, tutorial exposure, or a novelty effect.

Editorial extensions

If this is right

  • Chemists can specify precise modifications—keep this ring, replace this chloride with fluorine—by pointing at the molecule, removing the need to translate structural intent into prompt prose.
  • Candidates arrive with provenance (known compounds, safety signals, clinical phase, patents), so users can judge and justify AI suggestions in collaborative settings.
  • Branching history lets users revert, compare, and pursue multiple design directions, turning optimization from a linear sequence into an explorable graph.
  • Because the LLM only proposes edit plans and RDKit executes them, invalid or duplicate molecules are filtered before display; logs suggest roughly one valid, non-duplicate, constraint-satisfying candidate per two generation attempts.
  • Integrating property calculators and evidence lookup into one interface removes the repeated switching between separate generation, property, and visualization tools that participants described.

Reading between the lines

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

  • The same structure-as-interface pattern should transfer to other structure-intensive domains—protein engineering, materials discovery, reaction planning—where users can point at the object to constrain generation; the authors gesture at this but do not test it.
  • The reported gains may partly reflect novelty and tutorial warm-up rather than the interface alone; a component-level ablation would show whether anchor/avoid annotations, property sliders, or evidence cards each contribute independently.
  • The History Panel plus logged annotations is effectively an auto-generated design notebook; if it records decisions and evidence, it could make molecular optimization audits and reproducibility checks far cheaper than manual note-taking.
  • The edit-plan architecture is a reusable safety layer for LLM chemistry tools: restricting the model to executable operations and validating them deterministically bounds hallucination in a way that free-form molecule generation does not.
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. MolecularCanvas is an interactive system for LLM-assisted small-molecule optimization. It lets users specify high-level goals, structure-level annotations (anchor/avoid/free-form), property targets, and reference-based preferences; these are compiled into a structured intent representation, used to generate an explicit RDKit-executable edit plan, and candidates are validated, ranked, and presented with supporting evidence from ChEMBL/PubChem. The paper reports a formative study with six experts that yields five design requirements, the system design, and a user study with 12 participants in which MolecularCanvas is claimed to outperform a baseline on all 13 perceived-effectiveness items, SUS (72.5 vs. 60.8), and expert-rated final molecule quality (0.73 vs. 0.58). The authors also report pipeline reliability logs (95% executable edit plans, 57% valid candidates after filtering, 91% constraint satisfaction).

Significance. If the effectiveness claim were well supported, the paper would make a useful contribution: it addresses a real gap in generative molecular design tools by making structure-level intent explicit, providing evidence provenance, and integrating evaluation tools into one environment. The structured intent formalism and the two-stage edit-plan generation are sensible and reusable ideas, and the pipeline reliability logs are a valuable empirical addition. The formative study is appropriate for deriving design requirements. The main weakness is that the user-study evidence, which is load-bearing for the central claim, is confounded by the assignment of different tasks to the two conditions and by an uncontrolled baseline; the current data do not cleanly isolate the system effect.

major comments (3)
  1. [§5.1.4 / §5.2.1] The central effectiveness claim rests on a comparison that confounds system with task. Each participant used one system on Task A (nalidixic acid → quinolone) and the other system on Task B (cefazolin → cephalosporin), with only the order counterbalanced. The two tasks differ in starting molecule, optimization goal, and constraint complexity. Thus each paired difference on Q1–Q13, SUS, and expert ratings is a difference across two different tasks, not a difference between systems on the same task. If one task is intrinsically easier or more satisfying, the observed superiority of MolecularCanvas could arise from task assignment even with no true system benefit. The paper does not report per-task means or include task as a factor in any analysis. To support the claim, the authors must provide per-task descriptive statistics and either (a) compare systems within each task using the appropr
  2. [§5.1.2] The baseline condition is not a fixed comparator: 'participants were allowed to use any tools they were familiar with in the baseline conditions.' These tools vary per participant (different conversational AI, property calculators, editors), and the paper reports no measure of prior familiarity or fluency with the baseline tools or with MolecularCanvas after the tutorial. A participant who is less skilled with their self-selected baseline may rate it lower for reasons unrelated to MolecularCanvas, and a novelty or tutorial effect could inflate the test condition. The authors should either use a matched alternative interface or, at minimum, measure and statistically control for tool familiarity and prior experience, and report the robustness of the results under such controls.
  3. [§5.2.1] The quantitative analysis reports p-values for 13 questionnaire items and the SUS without multiple-comparison correction and without effect sizes or confidence intervals. With n=12 and a large number of significance tests, the reported 'all p<0.05' pattern is difficult to interpret. The authors should report exact test statistics (e.g., Wilcoxon W or t-value), effect sizes (e.g., rank-biserial or Cliff's delta), and either apply a correction or explicitly justify why it is unnecessary. This is also needed to help readers evaluate the magnitude of the perceived differences, which appear substantial on several items but are not quantified beyond means and standard deviations.
minor comments (5)
  1. [§4.3.2/§4.3.3] Near-duplicate removal uses Tanimoto similarity >0.85 and evidence retrieval uses ≥0.8, but these thresholds are not justified or varied in sensitivity analysis. Since they are free parameters of the pipeline, a brief rationale or robustness check would strengthen the presentation.
  2. [§5.2.1] The paper states 'Wilcoxon signed-rank test, p<0.001' for expert ratings and 'Wilcoxon p<0.05' for SUS, but the questionnaire items do not state which test was used. Please specify the test and whether it was paired or otherwise, and clarify how ties in Likert ratings were handled.
  3. [§5.1.4 / Figure 3] Figure 3 shows paired distributions for Q1–Q13, but individual-level paired data would be more informative, especially for judging the consistency and direction of the differences. Consider adding a paired-by-participant visualization or data table in the supplementary.
  4. [§5.2.1] The generation reliability numbers (95% executable plans, 57% valid candidates, 91% constraint satisfaction) are useful, but the 57% is the fraction passing validity and diversity filtering; the paper should state the base number of attempts and the number of candidates actually presented to users, so readers can gauge the practical user-facing yield.
  5. [§6.2] The limitations section is honest about missing ablations, but a sentence in Section 5.2.1 or Section 6 could note that the user study does not establish which components (e.g., anchor/avoid annotations, evidence panels, integrated property views) drive the observed differences. This would help set expectations for future work.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: MolecularCanvas is an empirical HCI systems evaluation; its effectiveness claim rests on a user study and external expert ratings, not on a self-derivation or fitted prediction.

full rationale

This is a systems paper, not a derivation. The central claim that MolecularCanvas improves molecular optimization outcomes is supported by a paired user study with 12 participants (Section 5.1.4) and by independent expert ratings of final molecules (Section 5.2.1: 'the molecules generated with MolecularCanvas received higher ratings than those from the baseline (mean 0.73 for MolecularCanvas vs. 0.58 for the baseline...)' with the expert evaluation anonymized and presented in random order). The computational pipeline (Section 4.3) is a prompt-and-validate pipeline around GPT-4o and RDKit; it fits no parameters to a target result and then re-presents that fit as a prediction. The reported generation reliability rates (95.0% executable edit plans, 57.0% validity/diversity pass, 91.0% structural-constraint satisfaction) are self-consistency checks on the generator, not claims that those same internally computed constraint scores independently validate the system's usefulness. The paper's few self-citations ([42], [43], [44], [47], [49], [56]) occur in related-work and future-work contexts and are not load-bearing for the main effectiveness claim. The strongest validity concerns are not circularity: the baseline was uncontrolled ('participants were allowed to use any tools they were familiar with in the baseline conditions', Section 5.1.2), and the two formal tasks were paired with the two systems in counterbalanced order (Section 5.1.4), so task difficulty may be confounded with system. Such a confound is an internal-validity limitation, not an equivalence-by-construction between conclusion and input. The paper's own limitation section honestly states that no component-level ablation was conducted (Section 6.2: 'We did not conduct a component-level ablation study...'), which acknowledges missing evidence rather than disguising a circular step. Overall, no step in the paper's argument reduces to its own inputs by definition, fitting, or self-citation chain.

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

MolecularCanvas introduces no scientific entities such as new particles or forces. Its load-bearing inputs are design choices (similarity thresholds, unspecified scoring weights, and top-k) plus domain assumptions about LLM edit-plan generation, RDKit execution, computational property prediction, and the validity of user-study measurements as evidence of real-world utility.

free parameters (4)
  • Near-duplicate Tanimoto threshold = 0.85
    Section 4.3.2: near-duplicates are removed when Morgan fingerprint Tanimoto coefficient > 0.85. This hand-chosen threshold controls diversity filtering and affects which candidates survive.
  • Evidence similarity threshold = 0.8
    Section 4.3.3: structurally similar known compounds are retrieved when Tanimoto similarity >= 0.8. This hand-chosen threshold determines which evidence is shown for each candidate.
  • Top-k candidate selection k = not specified
    Section 4.3.3: candidates are ranked and the top-k are selected for user exploration, but k is never reported, affecting the number of candidates users see.
  • Candidate scoring weights = not specified
    Section 4.3.3: property satisfaction, structural satisfaction, and chemical plausibility are aggregated into an overall score, but the weights used in the aggregation are not reported. Different weights would reorder the candidate list.
assumptions (4)
  • domain assumption GPT-4o can produce executable edit plans from structured molecular intent
    Section 4.3.2: the pipeline relies on the LLM to output add/remove/replace operations that RDKit can execute. If the model cannot reliably map constraints to valid operations, candidate quality degrades. The paper reports 95% executability on study logs, so this assumption is partially empirically supported but model-dependent.
  • domain assumption RDKit can faithfully execute the specified edit operations while preserving intended structure
    Section 4.3.2: replace/add/remove operations are applied via RDKit. This assumes chemical transformations are mechanically executable and that the 'largest remaining connected structure' heuristic preserves the user's intended scaffold.
  • domain assumption Property predictions from integrated computational tools are accurate enough for optimization decisions
    Sections 4.2.1 and 4.3.3: the system computes properties such as cLogP, TPSA, and drug-likeness heuristics and uses them to score candidates. If these predictions are inaccurate, user decisions and the candidate ranking are misleading.
  • domain assumption Self-reported ratings and two expert quality judgments are valid proxies for real-world molecular optimization success
    Section 5: the central effectiveness claim is based on 12 participants' Likert responses, SUS, and two external experts' ratings of final molecules. This assumes these measurements capture how well the tool would support real iterative drug discovery, which the paper does not directly validate.

how reviews work

0 comments
Cite this review

Pith. "Pith review of MolecularCanvas: LLM-assisted Small-Molecule Drug Discovery via Structure-Guided Constraints." pith.science (2026). https://pith.science/paper/HYHIEO3X

@misc{pith2026260800393,
  author       = {Pith},
  title        = {Pith review of: MolecularCanvas: LLM-assisted Small-Molecule Drug Discovery via Structure-Guided Constraints},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HYHIEO3X}},
  note         = {Machine review of arXiv:2608.00393}
}
read the original abstract

Small-molecule drug discovery relies on iterative molecular optimization, where chemists repeatedly modify candidate compounds to balance multiple competing properties such as efficacy, toxicity, and solubility. Recent advances in generative AI (GenAI) have shown promise in accelerating this process by automatically proposing new molecular structures or targeted modifications. However, existing GenAI-based molecular design tools remain poorly aligned with experts' real-world workflows. Specifically, they offer limited support for specifying structure-level modification intents on molecules, provide insufficient transparency into model-generated modifications, and lack integrated support for downstream property evaluation with external computational tools. To address these challenges, we introduce MolecularCanvas, an interactive system that enables users to iteratively construct an optimization context by integrating high-level goals, structure-level annotations, property constraints, and reference-based preferences. This context guides the generation of candidate molecules across diverse molecular structures. MolecularCanvas further enhances transparency by providing evidence for AI-generated suggestions and streamlines molecular evaluation by integrating commonly used computational tools for property assessment into a unified interface. Finally, a user study with 12 participants demonstrates the usefulness and effectiveness of MolecularCanvas in helping users optimize candidate molecules.

Figures

Figures reproduced from arXiv: 2608.00393 by the authors.

Figure 1
Figure 1. The interface of MolecularCanvas supports iterative molecular optimization through coordinated panels. (A) specifies high-level optimization goals and initializes the molecule. (B) enables structure-level editing with anchor, avoid, and in-context prompting to express modification intent. (C) supports property-oriented intent by exposing current values and target ranges. (D) externalizes and reuses reference-based s… view at source ↗
Figure 2
Figure 2. This specification contains two components. The first is [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Participants’ ratings for MolecularCanvas and the baseline across Q1–Q13. Each item shows paired distributions (top: MolecularCanvas, bottom: baseline). MolecularCanvas is rated significantly higher than the baseline on 13 questions (p < 0.05). (Q1, 4.17 ± 0.58 vs. 2.67 ± 1.30, 𝑝 < 0.01). We further organize the remaining results into four perspectives: modification intent expression, molecule generation quality, ev… view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

66 extracted references · 42 canonical work pages

  1. [1]

    Glassman

    Ian Arawjo, Chelse Swoopes, Priyan Vaithilingam, Martin Wattenberg, and Elena L. Glassman. 2024. ChainForge: A Visual Toolkit for Prompt Engineer- ing and LLM Hypothesis Testing. InProceedings of the 2024 CHI Conference on Human Factors in Computing Systems(Honolulu, HI, USA)(CHI ’24). Asso- ciation for Computing Machinery, New York, NY, USA, Article 304,...

    arXiv 2024

  2. [2]

    Hartmut Beck, Michael Härter, Bastian Haß, Carsten Schmeck, and Lars Baer- facker. 2022. Small Molecules and Their Impact in Drug Discovery: A Perspective on the Occasion of the 125th Anniversary of the Bayer Chemical Research Labo- ratory.Drug Discovery Today27, 6 (2022), 1560–1574. doi:10.1016/j.drudis.2022. 02.015

    work page doi:10.1016/j.drudis.2022 2022

  3. [3]

    Luca Beurer-Kellner, Marc Fischer, and Martin Vechev. 2023. Prompting Is Pro- gramming: A Query Language for Large Language Models.Proc. ACM Program. Lang.7, PLDI, Article 186 (June 2023), 24 pages. doi:10.1145/3591300

    doi:10.1145/3591300 2023

  4. [4]

    Richard Bickerton, Gaia V

    G. Richard Bickerton, Gaia V. Paolini, Jérémy Besnard, Sorel Muresan, and An- drew L. Hopkins. 2012. Quantifying the Chemical Beauty of Drugs.Nature Chemistry4, 2 (2012), 90–98. doi:10.1038/nchem.1243

    doi:10.1038/nchem.1243 2012

  5. [5]

    Andres M Bran, Sam Cox, Oliver Schilter, Carlo Baldassari, Andrew D White, and Philippe Schwaller. 2024. Augmenting Large Language Models with Chemistry Tools.Nature Machine Intelligence6, 5 (2024), 525–535. doi:10.1038/s42256-024- 00832-8

    doi:10.1038/s42256-024- 2024

  6. [6]

    Quick and Dirty

    John Brooke. 1996. SUS: A “Quick and Dirty” Usability Scale. InUsability Evalu- ation in Industry, Patrick W. Jordan, Bruce Thomas, Bernard A. Weerdmeester, and Ian L. McClelland (Eds.). Taylor & Francis, London, 189–194

    work page 1996

  7. [7]

    Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, Sandhini Agarwal, Ariel Herbert-Voss, Gretchen Krueger, Tom Henighan, Rewon Child, Aditya Ramesh, Daniel Ziegler, Jeffrey Wu, Clemens Winter, Chris Hesse, Mark Chen, Eric Sigler, Mateusz Litwin, Scott Gr...

    arXiv 2020

  8. [8]

    Zixin Chen, Yuhang Zeng, Sicheng Song, Yanna Lin, Xian Xu, Huamin Qu, and Meng Xia. 2026. VizQStudio: Iterative Visualization Literacy MCQs Design With Simulated Students.IEEE Transactions on Visualization and Computer Graphics 32, 8 (2026), 7468–7485. doi:10.1109/TVCG.2026.3695959

    work page arXiv 2026

Show all 66 references
  1. [9]

    DaEun Choi, Sumin Hong, Jeongeon Park, John Joon Young Chung, and Juho Kim

  2. [10]

    John Joon Young Chung and Eytan Adar. 2023. PromptPaint: Steering Text- to-Image Generation Through Paint Medium-like Interactions. InProceedings of the 36th Annual ACM Symposium on User Interface Software and Technology. Association for Computing Machinery, New York, NY, USA,...

    2023

  3. [11]

    Antoine Daina, Olivier Michielin, and Vincent Zoete. 2017. SwissADME: a free web tool to evaluate pharmacokinetics, drug-likeness and medicinal chemistry friendliness of small molecules.Scientific Reports7, 1 (2017), 42717. doi:10.1038/ srep42717

    2017

  4. [12]

    Carl Edwards, Tuan Lai, Kevin Ros, Garrett Honke, Kyunghyun Cho, and Heng Ji. 2022. Translation between Molecules and Natural Language. InProceedings of the 2022 Conference on Empirical Methods in Natural Language Processing. Association for Computational Linguistics, Abu Dhab...

    2022 doi

  5. [13]

    EPAM Systems. 2024. Ketcher: Open-source Web-based Chemical Structure Editor. https://github.com/epam/ketcher

    2024

  6. [14]

    Wenhao Gao, Tianfan Fu, Jimeng Sun, and Connor Coley. 2022. Sample Efficiency Matters: A Benchmark for Practical Molecular Optimization. InAdvances in Neural Information Processing Systems, Vol. 35. Curran Associates, Inc., Red Hook, NY, USA, 21342–21357. doi:10.52202/068431-1551

    2022 doi

  7. [15]

    Leach, Graeme R

    Ed Griffen, Andrew G. Leach, Graeme R. Robb, and Daniel J. Warner. 2011. Matched Molecular Pairs as a Medicinal Chemistry Tool.Journal of Medici- nal Chemistry54, 22 (2011), 7739–7750. doi:10.1021/jm200452d

    2011 doi

  8. [16]

    Hann and György M

    Michael M. Hann and György M. Keserü. 2012. Finding the Sweet Spot: the Role of Nature and Nurture in Medicinal Chemistry.Nature Reviews Drug Discovery 11, 5 (2012), 355–365. doi:10.1038/nrd3701

    2012 doi

  9. [17]

    JP Hughes, S Rees, SB Kalindjian, and KL Philpott. 2011. Principles of Early Drug Discovery.British Journal of Pharmacology162, 6 (2011), 1239–1249. doi:10.1111/ j.1476-5381.2010.01127.x

    2011

  10. [18]

    Ross Irwin, Spyridon Dimitriadis, Jiazhen He, and Esben Jannik Bjerrum. 2022. Chemformer: A Pre-trained Transformer for Computational Chemistry.Machine Learning: Science and Technology3, 1 (2022), 015022. doi:10.1088/2632-2153/ac3ffb

    2022 doi

  11. [19]

    Shoichi Ishida, Tomohiro Sato, Teruki Honma, and Kei Terayama. 2025. Large Language Models Open New Way of AI-Assisted Molecule Design for Chemists. Journal of Cheminformatics17, 1 (2025), 36. doi:10.1186/s13321-025-00984-8

    2025 doi

  12. [20]

    Jan H. Jensen. 2019. A Graph-based Genetic Algorithm and Generative Mod- el/Monte Carlo Tree Search for the Exploration of Chemical Space.Chemical Science10 (2019), 3567–3572. doi:10.1039/C8SC05372C

    2019 doi

  13. [21]

    Wengong Jin, Regina Barzilay, and Tommi Jaakkola. 2018. Junction Tree Vari- ational Autoencoder for Molecular Graph Generation. InProceedings of the 35th International Conference on Machine Learning, Vol. 80. PMLR, 2323–2332. https://proceedings.mlr.press/v80/jin18a.html

    2018

  14. [22]

    Hyeonah Kim, Minsu Kim, Sanghyeok Choi, and Jinkyoo Park. 2024. Genetic- guided GFlowNets for Sample Efficient Molecular Optimization. InAdvances in Neural Information Processing Systems. Curran Associates, Inc., Red Hook, NY, USA, 42618–42648. doi:10.52202/079017-1350

    2024 doi

  15. [23]

    Tae Soo Kim, Yoonjoo Lee, Minsuk Chang, and Juho Kim. 2023. Cells, Genera- tors, and Lenses: Design Framework for Object-Oriented Interaction with Large Language Models. InProceedings of the 36th Annual ACM Symposium on User Interface Software and Technology. Association for C...

    2023

  16. [24]

    Diederik P Kingma and Max Welling. 2013. Auto-Encoding Variational Bayes. arXiv:1312.6114 [stat.ML] https://arxiv.org/abs/1312.6114

    2013 arXiv

  17. [25]

    Jordy Homing Lam and Vsevolod Katritch. 2025. Navigating Structure-based Drug Discovery with Emerging Innovations in Physics- and Knowledge-based Approaches.npj Drug Discovery2, 1 (2025), 29. doi:10.1038/s44386-025-00031-4

    2025 doi

  18. [26]

    Greg Landrum et al . 2026. RDKit: Open-Source Cheminformatics Software. Zenodo. doi:10.5281/zenodo.21291217

    2026 doi

  19. [27]

    Yanna Lin, Haotian Li, Leni Yang, Aoyu Wu, and Huamin Qu. 2023. Inksight: Leveraging sketch interaction for documenting chart findings in computational notebooks.IEEE Transactions on Visualization and Computer Graphics30, 1 (2023), 944–954

    2023

  20. [28]

    Christopher A Lipinski, Franco Lombardo, Beryl W Dominy, and Paul J Feeney

  21. [29]

    Pengfei Liu, Weizhe Yuan, Jinlan Fu, Zhengbao Jiang, Hiroaki Hayashi, and Graham Neubig. 2023. Pre-train, Prompt, and Predict: A Systematic Survey of Prompting Methods in Natural Language Processing.ACM Comput. Surv.55, 9, Article 195 (Jan. 2023), 35 pages. doi:10.1145/3560815

    2023 doi

  22. [30]

    Loeffler, Jiazhen He, Alessandro Tibo, Jon Paul Janet, Alexey Voronov, Lewis H

    Hannes H. Loeffler, Jiazhen He, Alessandro Tibo, Jon Paul Janet, Alexey Voronov, Lewis H. Mervin, and Ola Engkvist. 2024. Reinvent 4: Modern AI–driven Generative Molecule Design.Journal of Cheminformatics16, 1 (2024), 20. doi:10.1186/s13321-024-00812-5

    2024 doi

  23. [31]

    Damien Masson, Sylvain Malacria, Géry Casiez, and Daniel Vogel. 2024. Direct- GPT: A Direct Manipulation Interface to Interact with Large Language Models. InProceedings of the 2024 CHI Conference on Human Factors in Computing Systems. Association for Computing Machinery, New Y...

    2024

  24. [32]

    Maurer, Martin Edwards, David Hepworth, Patrick Verhoest, and Charlotte M.N

    Tristan S. Maurer, Martin Edwards, David Hepworth, Patrick Verhoest, and Charlotte M.N. Allerton. 2022. Designing Small Molecules for Therapeutic Suc- cess: A Contemporary Perspective.Drug Discovery Today27, 2 (2022), 538–546. doi:10.1016/j.drudis.2021.09.017

    2022 doi

  25. [33]

    Janosch Menke, Yasmine Nahal, Esben Jannik Bjerrum, Mikhail Kabeshov, Samuel Kaski, and Ola Engkvist. 2024. Metis: A Python-based User Interface to Collect Expert Feedback for Generative Chemistry Models.Journal of Cheminformatics 16, 1 (2024), 100. doi:10.1186/s13321-024-00892-3

    2024 doi

  26. [34]

    Riza Özçelik, Helena Brinkmann, Emanuele Criscuolo, and Francesca Grisoni

  27. [35]

    Stephan Pajer, Marc Streit, Thomas Torsney-Weir, Florian Spechtenhauser, Torsten Möller, and Harald Piringer. 2017. WeightLifter: Visual Weight Space Ex- ploration for Multi-Criteria Decision Making.IEEE Transactions on Visualization and Computer Graphics23, 1 (2017), 611–620....

    2017

  28. [36]

    Gillet, Peter Willett, Visakan Kadirkamanathan, Christopher N

    George Papadatos, Muhammad Alkarouri, Valerie J. Gillet, Peter Willett, Visakan Kadirkamanathan, Christopher N. Luscombe, Gianpaolo Bravi, Nicola J. Rich- mond, Stephen D. Pickett, Jameed Hussain, John M. Pritchard, Anthony W. J. Cooper, and Simon J. F. Macdonald. 2010. Lead O...

    2010

  29. [37]

    Xiaohan Peng, Janin Koch, and Wendy E. Mackay. 2024. DesignPrompt: Using Multimodal Interaction for Design Exploration with Generative AI. InProceedings of the 2024 ACM Designing Interactive Systems Conference. Association for Com- puting Machinery, New York, NY, USA, 804–818....

    2024

  30. [38]

    Mihalovits, Maximilian Beckers, Nikolas Fechner, Niko- laus Stiefl, Finton Sirockin, William McCoull, Emma Evertsson, Malin Lemurell, Gergely Makara, and György M

    Anita Rácz, Levente M. Mihalovits, Maximilian Beckers, Nikolas Fechner, Niko- laus Stiefl, Finton Sirockin, William McCoull, Emma Evertsson, Malin Lemurell, Gergely Makara, and György M. Keserű. 2025. The Changing Landscape of Medicinal Chemistry Optimization.Nature Reviews Dr...

    2025 doi

  31. [39]

    Sadybekov and Vsevolod Katritch

    Anastasiia V. Sadybekov and Vsevolod Katritch. 2023. Computational Approaches Streamlining Drug Discovery.Nature616, 7958 (2023), 673–685. doi:10.1038/ s41586-023-05905-z

    2023

  32. [40]

    Jeff Sauro and James R. Lewis. 2016. Chapter 9 - Six enduring controversies in measurement and statistics. InQuantifying the User Experience (Second Edition) (second edition ed.), Jeff Sauro and James R. Lewis (Eds.). Morgan Kaufmann, Boston, 249–276. doi:10.1016/B978-0-12-802...

    2016 doi

  33. [41]

    Vaucher, Vishnu H

    Philippe Schwaller, Daniel Probst, Alain C. Vaucher, Vishnu H. Nair, David Kreutter, Teodoro Laino, and Jean-Louis Reymond. 2021. Mapping the Space of Chemical Reactions using Attention-based Neural Networks.Nature Machine Intelligence3, 2 (2021), 144–152. doi:10.1038/s42256-0...

    2021 doi

  34. [42]

    Rui Sheng, Chuhan Shi, Sobhan Lotfi, Shiyi Liu, Adam Perer, Huamin Qu, and Furui Cheng. 2026. Design patterns of human-AI interfaces in health- care.International Journal of Human-Computer Studies209 (2026), 103737. doi:10.1016/j.ijhcs.2026.103737

    2026

  35. [43]

    Rui Sheng, Xingbo Wang, Jiachen Wang, Xiaofu Jin, Zhonghua Sheng, Zhenxing Xu, Suraj Rajendran, Huamin Qu, and Fei Wang. 2026. TrialCompass: Visual Analytics for Enhancing the Eligibility Criteria Design of Clinical Trials.IEEE Transactions on Visualization and Computer Graphi...

    2026

  36. [44]

    Rui Sheng, Leni Yang, Haotian Li, Yan Luo, Ziyang Xu, Zhilan Zhou, David Gotz, and Huamin Qu. 2023. Knowledge Compass: A Question Answering System Guid- ing Students with Follow-Up Question Recommendations. InAdjunct Proceedings of the 36th Annual ACM Symposium on User Interfa...

    2023

  37. [45]

    Shrout and Joseph L

    Patrick E. Shrout and Joseph L. Fleiss. 1979. Intraclass correlations: uses in assessing rater reliability.Psychological Bulletin86, 2 (1979), 420–428. doi:10. 1037/0033-2909.86.2.420

    1979

  38. [46]

    Sicheng Song, Yanjie Zhang, Zixin Chen, Huamin Qu, Changbo Wang, and Chenhui Li. 2026. VizDefender: Unmasking Visualization Tampering Through Proactive Localization and Intent Inference.IEEE Transactions on Visualization and Computer Graphics32, 6 (2026), 4720–4730. doi:10.110...

    2026

  39. [47]

    Sicheng Song, Yipeng Zhang, Yanna Lin, Huamin Qu, Changbo Wang, and Chen- hui Li. 2025. GVVST: Image-Driven Style Extraction From Graph Visualizations for Visual Style Transfer.IEEE Transactions on Visualization and Computer Graph- ics31, 9 (2025), 5975–5989. doi:10.1109/TVCG....

    2025

  40. [48]

    Stokes, Kevin Yang, Kyle Swanson, Wengong Jin, Andres Cubillos- Ruiz, Nina M

    Jonathan M. Stokes, Kevin Yang, Kyle Swanson, Wengong Jin, Andres Cubillos- Ruiz, Nina M. Donghia, Craig R. MacNair, Shawn French, Lindsey A. Carfrae, Zohar Bloom-Ackermann, Victoria M. Tran, Anush Chiappino-Pepe, Ahmed H. Badran, Ian W. Andrews, Emma J. Chory, George M. Churc...

    2020 doi

  41. [49]

    Yushi Sun, Kai Sun, Yifan Ethan Xu, Xiao Yang, Xin Luna Dong, Nan Tang, and Lei Chen. 2025. KERAG: Knowledge-Enhanced Retrieval-Augmented Generation for Advanced Question Answering. InFindings of the Association for Computational Linguistics: EMNLP 2025. Association for Comput...

    2025 doi

  42. [50]

    Siraki, and Khaled Barakat

    Yasser Tabana, Dinesh Babu, Richard Fahlman, Arno G. Siraki, and Khaled Barakat

  43. [51]

    Daniel F Veber, Stephen R Johnson, Hung-Yuan Cheng, Brian R Smith, Keith W Ward, and Kenneth D Kopple. 2002. Molecular Properties that Influence the Oral Bioavailability of Drug Candidates.Journal of medicinal chemistry45, 12 (2002), 2615–2623. doi:10.1021/jm020017n

    2002 doi

  44. [52]

    Waring, John Arrowsmith, Andrew R

    Michael J. Waring, John Arrowsmith, Andrew R. Leach, Paul D. Leeson, Sam Mandrell, Robert M. Owen, Garry Pairaudeau, William D. Pennie, Stephen D. Pickett, Jibo Wang, Owen Wallace, and Alex Weir. 2015. An Analysis of the Attrition of Drug Candidates from Four Major Pharmaceuti...

    2015 doi

  45. [53]

    Tongshuang Wu, Michael Terry, and Carrie Jun Cai. 2022. AI Chains: Transparent and Controllable Human-AI Interaction by Chaining Large Language Model Prompts. InProceedings of the 2022 CHI Conference on Human Factors in Computing Systems. Association for Computing Machinery, N...

    2022

  46. [54]

    Zhenqin Wu, Bharath Ramsundar, Evan N Feinberg, Joseph Gomes, Caleb Ge- niesse, Aneesh Pappu, Karl Leswing, and Vijay Pande. 2018. MoleculeNet: A Benchmark for Molecular Machine Learning.Chemical Science9 (2018), 513–530. doi:10.1039/C7SC02664A

    2018 doi

  47. [55]

    Guoli Xiong, Zhenxing Wu, Jiacai Yi, Li Fu, Zhijiang Yang, Changyu Hsieh, Mingzhu Yin, Xiangxiang Zeng, Chengkun Wu, Aiping Lu, et al. 2021. ADMETlab 2.0: An Integrated Online Platform for Accurate and Comprehensive Predictions of ADMET Properties.Nucleic acids research49, W1 ...

    2021

  48. [56]

    Xiao Yang, Kai Sun, Hao Xin, Yushi Sun, Nikita Bhalla, Xiangsen Chen, Sajal Choudhary, Rongze Daniel Gui, Ziran Will Jiang, Ziyu Jiang, Lingkun Kong, Brian Moran, Jiaqi Wang, Yifan Ethan Xu, An Yan, Chenyu Yang, Eting Yuan, Hanwen Zha, Nan Tang, Lei Chen, Nicolas Scheffer, Yue...

    2024

  49. [57]

    Zamfirescu-Pereira, Richmond Y

    J.D. Zamfirescu-Pereira, Richmond Y. Wong, Bjoern Hartmann, and Qian Yang

  50. [58]

    Odin Zhang, Haitao Lin, Hui Zhang, Huifeng Zhao, Yufei Huang, Chang-Yu Hsieh, Peichen Pan, and Tingjun Hou. 2024. Deep Lead Optimization: Leveraging Generative AI for Structural Modification.Journal of the American Chemical Society146, 46 (2024), 31357–31370. doi:10.1021/jacs.4c11686

    2024 doi

  51. [59]

    Yang Zhang, Jie Li, and Xu Chao. 2024. ChemNav: An Interactive Visual Tool to Navigate in the Latent Space for Chemical Molecules Discovery.Visual Informatics 8, 4 (2024), 60–70. doi:10.1016/j.visinf.2024.10.002

    2024 doi

  52. [60]

    Wanjie Zheng, Jie Li, and Yang Zhang. 2023. Desirable Molecule Discovery via Generative Latent Space Exploration.Visual Informatics7, 4 (2023), 13–21. doi:10.1016/j.visinf.2023.10.002

    2023 doi

  53. [61]

    Kaijie Zhu, Qinlin Zhao, Hao Chen, Jindong Wang, and Xing Xie. 2024. Prompt- Bench: A Unified Library for Evaluation of Large Language Models.Journal of Machine Learning Research25, 254 (2024), 1–22. https://arxiv.org/abs/2312.07910

    2024 arXiv

  54. [62]

    InProceedings of the 2023 CHI Conference on Human Factors in Computing Systems

    Why Johnny Can’t Prompt: How Non-AI Experts Try (and Fail) to Design LLM Prompts. InProceedings of the 2023 CHI Conference on Human Factors in Computing Systems. Association for Computing Machinery, New York, NY, USA, Article 437, 21 pages. doi:10.1145/3544548.3581388

    2023

  55. [2012]

    doi:10.1016/j.addr.2012.09.019

    Experimental and Computational Approaches to Estimate Solubility and Permeability in Drug Discovery and Development Settings.Advanced drug delivery reviews64 (2012), 4–17. doi:10.1016/j.addr.2012.09.019

    2012 doi

  56. [2023]

    doi:10.1186/ s12896-023-00815-4

    Target Identification of Small Molecules: An Overview of the Current Applications in Drug Discovery.BMC Biotechnology23, 1 (2023), 44. doi:10.1186/ s12896-023-00815-4

    2023

  57. [2024]

    InProceedings of the 2024 CHI Conference on Human Factors in Computing Systems (CHI ’24)

    CreativeConnect: Supporting Reference Recombination for Graphic Design Ideation with Generative AI. InProceedings of the 2024 CHI Conference on Human Factors in Computing Systems (CHI ’24). Association for Computing Machinery, New York, NY, USA, Article 1055, 25 pages. doi:10....

    2024

  58. [2025]

    doi:10.1021/acs.jcim.5c00641

    Generative Deep Learning for de Novo Drug Design-A Chemical Space Odyssey.Journal of Chemical Information and Modeling65, 14 (2025), 7352–7372. doi:10.1021/acs.jcim.5c00641

    2025 doi

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.