Writing study: Video-Based Optimal Transport for Feedback-Efficient Offline Preference-Based Reinforcement Learning

ICML overviewCorpus atlasBy contribution typeVisual galleryRules & templateWriting patternsPaper studies

← ICML 2026 oral writing corpus

Tung Minh Luu, Hwanhee Kim, Younghwan Lee, Chang D. Yoo · 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: Training, objectives & optimization (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

23 physical pages

Before references / appendix

≈ 8.36 page-equivalents

Main figures / tables

7 / 3

Attached appendix figures / tables

12 / 10

Attached appendix span

≈ 9.91 page-equivalents

Before references / appendix: ≈ 36.35%References / intervening material: ≈ 20.57%Attached appendix: ≈ 43.09%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	≈ 0.9 pp.	3.91%
2. Related Work	2	≈ 0.56 pp.	2.43%
3. Preliminaries	2	≈ 0.75 pp.	3.26%
4. Method	3	≈ 1.54 pp.	6.7%
5. Experiments	5	≈ 3.83 pp.	16.65%
6. Discussion	8	≈ 0.3 pp.	1.3%
Acknowledgements	9	≈ 0.1 pp.	0.43%
Impact Statement	9	≈ 0.05 pp.	0.22%
References	9	≈ 4.73 pp.	20.57%
Appendix	14	≈ 0.22 pp.	0.96%
A. Details on Experiments	14	≈ 1.5 pp.	6.52%
B. Learning Curves	15	≈ 3.27 pp.	14.22%
C. Additional Results and Analysis	19	≈ 2 pp.	8.7%
D. Extended Discussion on Broader Context and Future Directions	21	≈ 2.91 pp.	12.65%

All detected figure and table locations

Main text

• Figure 1 · p. 4
• Table 1 · p. 6
• Figure 2 · p. 6
• Figure 3 · p. 6
• Figure 4 · p. 7
• Figure 5 · p. 7
• Figure 6 · p. 7
• Table 2 · p. 8
• Figure 7 · p. 8
• Table 3 · p. 8

Attached appendix

• Figure 8 · p. 14
• Table 4 · p. 15
• Table 5 · p. 15
• Table 6 · p. 15
• Figure 9 · p. 15
• Table 7 · p. 16
• Table 8 · p. 16
• Figure 10 · p. 16
• Figure 11 · p. 16
• Figure 12 · p. 17
• Figure 13 · p. 17
• Figure 14 · p. 18
• Figure 15 · p. 18
• Figure 16 · p. 18
• Table 9 · p. 19
• Table 10 · p. 19
• Table 11 · p. 19
• Table 12 · p. 20
• Table 13 · p. 20
• Figure 17 · p. 22
• Figure 18 · p. 23
• Figure 19 · p. 23

Argument and organization

1. 1 Introduction; 2 Related Work; 3 Preliminaries

   Defines the feedback-cost bottleneck and preference/transport notation.

2. 4 Method

   Moves from trajectory encoding to transported pseudo-preferences and a practical solver.

3. 5 Experiments; 6 Discussion

   Uses explicit research questions to organize benchmarks, ablations, nuisance tests and robot evaluation.

Problem, claim and contribution

The introduction connects annotation cost to the underused visual structure of trajectory preferences, making label efficiency the primary claim.

Abstract and Introduction, p.1 · Main paper

Argument and section order

The experiment opening lists five questions that map to low-label performance, components, hyperparameters, distractions and real robots.

Section 5 opening, p.5 · Main paper

Figures, captions and visual explanation

The overview pairs a trajectory-to-latent pipeline with a small numerical preference matrix and transport plan, connecting the conceptual flow to the score calculation.

Figure 1, p.4 · Main paper

Tables, comparisons and evidence

The main table defines bold as within five percent of the best non-oracle method rather than exact maximum; accompanying curves state five-run means and standard-deviation shading.

Table 1 and Figure 2, p.6 · Main paper

Appendix and reproducibility

Supplementary human-teacher comparisons and pseudo-label accuracy are separated from the main scripted-teacher benchmark, with the excluded tied-label condition stated.

Appendix C.4-C.5, p.20; Algorithm 1 and data details, p.14 · Main paper

Clarity and writing conventions

The main setup discloses synthetic preference teachers and an exception using human labels; the real-robot results explicitly use ten evaluation episodes.

Section 5.1, p.5; Section 5.5 and Table 3, p.8 · Main paper

What to reuse—and what not to infer

Transferable lesson: Map experimental questions to evidence blocks and disclose label provenance and display conventions where the reader encounters the numbers.

Boundary: Most benchmark preferences are scripted, robot evidence is small, and visual-encoder biases and transport scaling remain stated limitations.

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, 5, 7, 8, 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 14, 20. Remaining appendix proofs and details were not exhaustively read. No separately linked supplement inspected.

Visually inspected PDF pages

4, 6

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 · 23 pages