Writing study: Incentivizing Truthfulness and Collaborative Fairness in Bayesian Learning

ICML overviewCorpus atlasBy contribution typeVisual galleryRules & templateWriting patternsPaper studies

← ICML 2026 oral writing corpus

Rachael Hwee Ling Sim, Jue Fan, Xiao Tian, Xinyi Xu, Patrick Jaillet, Bryan Kian Hsiang Low · 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: Theory & formal analysis (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

38 physical pages

Before references / appendix

≈ 9.18 page-equivalents

Main figures / tables

3 / 1

Attached appendix figures / tables

16 / 3

Attached appendix span

≈ 26.91 page-equivalents

Before references / appendix: ≈ 24.16%References / intervening material: ≈ 5.03%Attached appendix: ≈ 70.82%Estimated layout envelopes, not text/artwork area. Unresolved appendix boundaries remain within the undivided post-body material.

Source-bound correction: Replace a prose cross-reference falsely detected as a caption with the actual caption from physical-page source text and coordinates.

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.	3.58%
2. Collaborative ML	2	≈ 0.91 pp.	2.39%
3. Data Valuation Functions	3	≈ 1.21 pp.	3.18%
4. Fair and Truthful Semivalues	4	≈ 0.86 pp.	2.26%
5. Other Incentives or Constraints	5	≈ 1.03 pp.	2.71%
6. Experiments	6	≈ 1.93 pp.	5.08%
7. Related Works	8	≈ 0.41 pp.	1.08%
8. Conclusion	9	≈ 1 pp.	2.63%
Impact Statement	10	≈ 0.05 pp.	0.13%
Acknowledgments	10	≈ 0.08 pp.	0.21%
References	10	≈ 1.91 pp.	5.03%
A. Assumptions	12	≈ 1.07 pp.	2.82%
B. Background	13	≈ 6.46 pp.	17%
C. Data Valuation Functions	19	≈ 2.9 pp.	7.63%
D. Truthful Semivalues	22	≈ 1.15 pp.	3.03%
E. Considering Other Constraints	23	≈ 2.73 pp.	7.18%
F. Experiments	26	≈ 8.04 pp.	21.16%
G. Limitations	34	≈ 2.07 pp.	5.45%
H. Other Questions	36	≈ 2.5 pp.	6.58%

All detected figure and table locations

Main text

• Table 1 · p. 2
• Figure 1 · p. 7
• Figure 2 · p. 7
• Figure 3 · p. 8

Attached appendix

• Figure 4 · p. 14
• Table 2 · p. 15
• Figure 5 · p. 25
• Figure 6 · p. 27
• Figure 7 · p. 27
• Figure 8 · p. 28
• Figure 9 · p. 28
• Figure 10 · p. 29
• Figure 11 · p. 30
• Figure 12 · p. 30
• Figure 13 · p. 31
• Figure 14 · p. 31
• Figure 15 · p. 32
• Table 3 · p. 32
• Table 4 · p. 33
• Figure 16 · p. 33
• Figure 17 · p. 34
• Figure 18 · p. 35
• Figure 19 · p. 36

Argument and organization

1. 1 Introduction; 2 Collaborative ML; 3 Data Valuation

   Defines separate fairness and truthfulness requirements before constructing the valuation function.

2. 4 Fair and Truthful Semivalues; 5 Other Constraints

   Extends the guarantee to coalition rewards and examines budget and validation-data relaxations.

3. 6 Experiments; 7 Related Work; 8 Conclusion

   Tests the assumption-sensitive mechanism and directs readers to dedicated limitations and questions.

Problem, claim and contribution

The introduction presents fairness and truthful submission as distinct requirements and motivates why rewarding submitted data as-is can fail.

Abstract and Introduction, p.1 · Main paper

Argument and section order

Definitions precede the propositions; the coalition result explicitly adds an assumption about other sources rather than silently transferring the single-source guarantee.

Definitions 2.1 and 3.1, p.3; Proposition 3.2, p.4 · Main paper

Figures, captions and visual explanation

The experiments separate validation-size/noise curves from coalition reward plots, with repeated strategy encodings and captions specifying averaging and confidence intervals.

Figures 1-2, p.7 · Main paper

Tables, comparisons and evidence

The experimental roadmap states what is being validated and declines to claim superiority over accuracy-based valuation methods lacking the target theoretical guarantees.

Section 6, p.6 · Main paper

Appendix and reproducibility

A dedicated assumptions appendix explains no-cloning, shared Bayesian models and validation beliefs; a separate limitations section revisits validation quality and restriction to Bayesian models.

Appendices A and G, pp.12,34 · Main paper

Clarity and writing conventions

The main text explicitly reports that sufficiently misspecified noisy validation data can make untruthful strategies more valuable, preserving the difference between an equilibrium guarantee and robustness everywhere.

Section 6.1, p.7; budget discussion, p.5 · Main paper

What to reuse—and what not to infer

Transferable lesson: Name the desired properties separately, show which extra assumptions each extension needs, and test the boundary where the guarantee ceases to describe behavior.

Boundary: Truthfulness is an expectation/equilibrium statement under Bayesian and information assumptions, not a universal data-integrity guarantee; proofs were not independently verified.

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-oriented scan of the main argument through physical p.9, with focused physical-page text inspection on 1, 3, 4, 5, 6, 7, 9. Framing, section progression, mechanism/proof explanation, displays, comparison conditions and conclusions were inspected. This is not exhaustive technical reading, proof verification or reproduction.

Appendix / supplement

Appendix organization and main-text pointers were checked; focused text inspection of physical pages 12, 34. Remaining appendix proofs and details were not exhaustively read. No separately linked supplement inspected.

Visually inspected PDF pages

4, 7

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 proceedings PDF, PMLR volume 306, retrieved 2026-10-07 · 38 pages