Writing study: From Feasible to Practical: Pareto-Optimal Synthesis Planning

ICML overviewCorpus atlasBy contribution typeVisual galleryRules & templateWriting patternsPaper studies

← ICML 2026 oral writing corpus

Friedrich Hastedt, Dongda Zhang, Antonio Del Rio Chanona · Official ICML 2026 oral record

Observed writing choices, not acceptance rules. This study examines the inspected manuscript version. It does not establish why the paper was accepted or that its presentation choices should be copied unchanged.

Figure, table and section ledger

Editorial contribution type: Domain-grounded methods & physical systems (primary).

Layout extraction, not a required paper template. Counts are distinct caption labels; a multi-panel figure counts once. Section spans are estimated from heading positions, including floats and whitespace. Method and limitations.

Canonical PDF

21 physical pages

Before references / appendix

≈ 9.04 page-equivalents

Main figures / tables

5 / 4

Attached appendix figures / tables

1 / 7

Attached appendix span

≈ 8.91 page-equivalents

Before references / appendix: ≈ 43.05%References / intervening material: ≈ 14.52%Attached appendix: ≈ 42.43%Estimated layout envelopes, not text/artwork area. Unresolved appendix boundaries remain within the undivided post-body material.

Section-by-section lengths and source pages

The title/abstract/front matter precede the first numbered heading. A span ends at the next detected top-level heading, not at the last sentence; end-page whitespace is included. References and statements are listed separately. These are not measurements of text-only or artwork area.

Top-level heading	Starts on PDF page	Estimated span	Share of entire PDF
1. Introduction	1	≈ 1.36 pp.	6.48%
2. Background	2	≈ 1.17 pp.	5.57%
3. Methods	3	≈ 3.02 pp.	14.38%
4. Experiments	6	≈ 2.46 pp.	11.71%
5. Limitations and Future Outlook	9	≈ 0.32 pp.	1.52%
6. Conclusion	9	≈ 0.16 pp.	0.76%
Acknowledgements	9	≈ 0.09 pp.	0.43%
Impact Statement	9	≈ 0.14 pp.	0.67%
References	10	≈ 3.04 pp.	14.48%
A. Single-Step Retrosynthesis and Condition Prediction Models	13	≈ 0.31 pp.	1.48%
B. Proof of Theorem 3.4	13	≈ 1.18 pp.	5.62%
C. Objective Functions	14	≈ 1.86 pp.	8.86%
D. Search Hyperparameters	16	≈ 1.65 pp.	7.86%
E. Supplementary Results	18	≈ 3 pp.	14.29%
F. Toxicity Assessment Prompt Template	21	≈ 0.91 pp.	4.33%

All detected figure and table locations

Main text

• Figure 1 · p. 2
• Figure 2 · p. 4
• Figure 3 · p. 5
• Table 1 · p. 7
• Figure 4 · p. 7
• Figure 5 · p. 8
• Table 2 · p. 8
• Table 3 · p. 8
• Table 4 · p. 9

Attached appendix

• Table 5 · p. 18
• Table 6 · p. 18
• Table 7 · p. 18
• Figure 6 · p. 18
• Table 8 · p. 19
• Table 9 · p. 19
• Table 10 · p. 20
• Table 11 · p. 20

Argument and organization

1. Problem and contribution

   Moves from finding any route to exposing tradeoffs among routes, while making the single-step model and predicted reaction conditions explicit dependencies.

2. Argument sequence

   Graph values and batched search precede weight sampling and objective definitions; evaluation is split into Pareto quality, route success and chemical diversity.

3. Evidence and verification

   Per-molecule paired tests and dominance in both directions accompany hypervolume summaries; success-rate tables retain cases where the single-objective baseline wins.

Problem, claim and contribution

Moves from finding any route to exposing tradeoffs among routes, while making the single-step model and predicted reaction conditions explicit dependencies.

Introduction and contributions, pp. 1-2 · Main paper

Argument and section order

Graph values and batched search precede weight sampling and objective definitions; evaluation is split into Pareto quality, route success and chemical diversity.

Sections 3-4, pp. 3-8 · Main paper

Figures, captions and visual explanation

Pseudocode annotates selection, prediction and batched operations; the result page juxtaposes front-quality distributions with the modest loss in single-route success.

Algorithm 1 and Figure 3, p. 5; Figure 5 and Tables 2-3, p. 8; both PNGs inspected · Main paper

Tables, comparisons and evidence

Per-molecule paired tests and dominance in both directions accompany hypervolume summaries; success-rate tables retain cases where the single-objective baseline wins.

Section 4.2 and Tables 2-3, p. 8 · Main paper

Appendix and reproducibility

The proof states admissible-bound and termination conditions separately from practical budgets, and objective/prompt documentation exposes how toxicity proxies enter the search.

Appendix B, pp. 13-14; C, pp. 14-15; F, p. 21 · Main paper

Clarity and writing conventions

Numbered evaluation questions give different notions of practical utility separate evidence destinations rather than compressing them into a single score.

Section 4.2, pp. 7-8 · Main paper

What to reuse—and what not to infer

Transferable lesson: For domain planning, pair improved multi-objective quality with coverage losses and disclose how learned or judged objective proxies are constructed.

Boundary: A conditional Pareto guarantee for the specified search model is not a certificate under a finite time budget, and computed routes or toxicity scores are not experimental chemistry validation.

Apply this lesson using the argument and evidence worksheet.

Inspection record and sources

Checked

2026-10-07

Reading status

Writing-focused manuscript scan + selected visuals inspected

Main text

Writing-focused distributed manuscript scan, not exhaustive technical/proof validation. Sections 1-6, pp. 1-9: practical multi-objective synthesis, search representation, batched algorithm, weight strategies, conditional optimality and quality/success/diversity tests.

Appendix / supplement

Writing-focused appendix scan; code, data and external assets not executed. Appendices A-F, pp. 13-21: predictor descriptions, theorem proof, objectives, search settings, additional datasets/baselines and toxicity scoring prompt.

Visually inspected PDF pages

5, 8

This is a writing-focused scan of the manuscript and the stated supplementary scope: framing, organization, visual explanation, evidence presentation and wording. It is not a line-by-line proof audit, reproduction of results or exhaustive scientific review.

• Published ICML 2026 proceedings PDF, PMLR volume 306; retrieved 2026-10-07. Public OpenReview decision notes were not independently reverified. · 21 pages