Writing study: Moto: Latent Motion Token as the Bridging Language for Learning Robot Manipulation from Videos

ICCV overviewCorpus atlasBy contribution typeVisual galleryRules & templateWriting patternsPaper studies

← ICCV 2025 oral writing corpus

Yi Chen, Yuying Ge, Weiliang Tang, Yizhuo Li, Yixiao Ge, Mingyu Ding, Ying Shan, Xihui Liu · Official ICCV 2025 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

12 physical pages

Before references / appendix

≈ 8.09 page-equivalents

Main figures / tables

12 / 3

Attached appendix figures / tables

0 / 0

Attached appendix span

No boundary detected; not assumed zero

Before references / appendix: ≈ 67.42%References / intervening material: ≈ 32.58%Estimated layout envelopes, not text/artwork area. Unresolved appendix boundaries remain within the undivided post-body material.

Extraction notes: No attached appendix boundary detected. Zero means no recognized caption, not independently verified absence.

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	2	≈ 0.82 pp.	6.83%
2. Related Work	2	≈ 0.82 pp.	6.83%
3. Methodology	3	≈ 1.21 pp.	10.08%
4. Benchmarks and Datasets	4	≈ 0.91 pp.	7.58%
5. Experiments	5	≈ 3.04 pp.	25.33%
6. Conclusion and Discussion	8	≈ 0.2 pp.	1.67%
Acknowledgement	9	≈ 0.04 pp.	0.33%
References	9	≈ 3.91 pp.	32.58%

All detected figure and table locations

Main text

• Figure 1 · p. 1
• Figure 2 · p. 3
• Figure 3 · p. 4
• Figure 4 · p. 5
• Figure 5 · p. 5
• Figure 6 · p. 6
• Table 1 · p. 6
• Figure 7 · p. 6
• Table 2 · p. 7
• Table 3 · p. 7
• Figure 8 · p. 7
• Figure 9 · p. 7
• Figure 10 · p. 8
• Figure 11 · p. 8
• Figure 12 · p. 8

Attached appendix

No recognized caption labels.

Argument and organization

1. 1 Introduction; 2 Related Work

   Asks whether motion-focused video tokens can transfer without action annotations during pretraining.

2. 3 Methodology; 4 Benchmarks and Datasets

   Separates tokenizer learning, instruction-conditioned motion pretraining and action-supervised co-fine-tuning.

3. 5 Experiments; 6 Conclusion and Discussion

   Answers interpretability, learned-prior and control-performance questions, then studies data efficiency and motion-loss retention.

Problem, claim and contribution

A bridging motion language is positioned between video observation and embodiment-specific control, with the remaining action-adaptation gap stated explicitly.

Sections 1 and 3.4, physical pp.2,4 · Main paper

Argument and section order

Three experimental questions mirror tokenizer semantics, pretraining priors and downstream policy utility rather than collapsing them into one benchmark score.

Section 5, p.5 · Main paper

Figures, captions and visual explanation

The three-stage overview marks frozen encoders/tokenizer, inserted action queries and the retained motion-token objective, making supervision transitions visible.

Figure 2, p.3 · Main paper

Tables, comparisons and evidence

CALVIN results expose each method’s observation space and define chain length over 1,000 trials; SIMPLER retains task-specific non-wins despite the best aggregate.

Tables 2–3, p.7 · Main paper

Appendix and reproducibility

Implementation, benchmark settings, baseline details and additional generated trajectories are routed to supplementary material, not inspected here.

Sections 3.1,4,5.2–5.3, pp.3–4,6–7 · Main paper

Clarity and writing conventions

The human-video extension is called a preliminary study and separately reports the added 105k SSV2 videos alongside robot-video pretraining.

Section 5.3, pp.7–8 · Main paper

What to reuse—and what not to infer

Transferable lesson: Make every change in supervision and observation privileges visible across pretraining, fine-tuning and evaluation.

Boundary: Action-free pretraining is not action-free learning: downstream policy training uses labeled expert trajectories. Token visual consistency and trajectory likelihood are not proofs of causal motion understanding or calibrated rewards; small real-world task trials and selected human-video transfer do not establish general-purpose safe control.

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 across the main manuscript, physical PDF pp.1–8, covering framing, section progression, method explanations, captions, comparison protocols and concluding scope; reference boundary checked. Selected full-page rasters were inspected. Not exhaustive line-by-line reading, proof verification or experimental reproduction.

Appendix / supplement

Main-paper appendix/supplement references were checked. A separate supplementary source is linked in the inventory but was not retrieved or inspected; its contents are not included in the reading or visual-page counts.

Visually inspected PDF pages

3, 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.

• Official CVF open-access ICCV 2025 accepted-paper PDF, retrieved 7 October 2026; CVF states identical to accepted version except watermark; not assumed identical to initial submission · 12 pages