Writing study: SocialNav: Training Human-Inspired Foundation Model for Socially-Aware Embodied Navigation

CVPR overviewCorpus atlasBy contribution typeVisual galleryRules & templateWriting patternsPaper studies

← CVPR 2026 oral writing corpus

Ziyi Chen, Yingnan Guo, Zedong Chu, Minghua Luo, Yanfen Shen, Mingchao Sun, Junjun Hu, Shichao Xie, Yang Kuan, Pei Shi, Zhining Gu, Lu Liu, Honglin Han, Xiaolong Wu, Mu Xu, Yu Zhang · Official CVPR 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: Architecture & representation (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

11 physical pages

Before references / appendix

≈ 8.11 page-equivalents

Main figures / tables

4 / 4

Attached appendix figures / tables

0 / 0

Attached appendix span

Separate file recorded below; not in this PDF total

Before references / appendix: ≈ 73.73%References / intervening material: ≈ 26.27%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	1	≈ 1.18 pp.	10.73%
2. Related Work	2	≈ 0.28 pp.	2.55%
3. SocNav Dataset and Benchmark	2	≈ 1.92 pp.	17.45%
4. SocialNav Foundation Model	4	≈ 1.49 pp.	13.55%
5. Experiment	6	≈ 2.56 pp.	23.27%
6. Conclusion	8	≈ 0.28 pp.	2.55%
References	9	≈ 2.89 pp.	26.27%

All detected figure and table locations

Main text

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

Attached appendix

No recognized caption labels.

Separate source files—excluded from canonical-PDF totals

A separate author version may repeat main text and references. Whole-file lengths and counts are not extra appendix material and are never added to the canonical PDF totals. Read the study's source scope for which sections received substantive reading versus layout inspection.

Additional source 1

Separate PDF (or ZIP containing it): whole-file layout of 8 physical pages; 5 detected figures and 5 tables. These may include overlapping main-text material.

Located appendix only: starts on physical p.1; ≈ 7.93 page-equivalents, 5 figure labels and 5 table labels. This interval is distinct from the whole file.

Detected heading	PDF page	Estimated span
Supplementary Material	1	≈ 0.03 pp.
A. Supplementary Materials for the SocNav Dataset	1	≈ 0.16 pp.
B. Supplementary Materials for the SocialNav Foun-	1	≈ 0.05 pp.
C. Additional Experimental Results	1	≈ 7.68 pp.

• Figure 5 · p. 2
• Figure 6 · p. 3
• Figure 7 · p. 4
• Figure 8 · p. 4
• Table 5 · p. 5
• Table 6 · p. 6
• Table 7 · p. 6
• Table 8 · p. 7
• Table 9 · p. 7
• Figure 9 · p. 8

Argument and organization

1. Dataset and benchmark

   Define social supervision and evaluation conditions before the foundation-model design.

2. Hierarchy and staged training

   Connect high-level reasoning to low-level trajectories, then separate open-loop, closed-loop and real deployment evidence.

Problem, claim and contribution

The problem distinguishes socially acceptable paths from merely short or collision-free routes, giving social compliance its own evidence requirement.

Introduction and Figure 1 · Main paper · Additional source 2

Argument and section order

Data/benchmark construction comes before model/training sections, and open-loop, closed-loop and physical deployment results are separately titled.

Sections 3–5 · Main paper · Additional source 2

Figures, captions and visual explanation

Figure 3 combines a three-stage training strip with labeled Brain/Action Expert modules; Figure 4 aligns overhead routes with egocentric sequences and marks problematic baseline behavior.

Figure 3, PDF p. 4; Figure 4, p. 7 · Main paper · Additional source 2

Tables, comparisons and evidence

The CityWalker table explicitly distinguishes scenario-averaged Mean from sample-averaged All; real-world results report success counts by environment.

Table 1, PDF p. 6; Table 3, p. 7 · Main paper · Additional source 2

Appendix and reproducibility

The supplement makes prompts, trainable-module stages and reward-component ablations inspectable, rather than relying on a single fluent reasoning example.

Supplement Figures 7–8 and Tables 5–9, pp. 4–7 · Main paper · Additional source 2

Clarity and writing conventions

The conclusion acknowledges hand-crafted rewards and the limits of semantic traversability as a representation of broader conventions.

Conclusion, PDF p. 8 · Main paper · Additional source 2

What to reuse—and what not to infer

Transferable lesson: For embodied systems, separate prediction quality, closed-loop behavior and real-world success; label both aggregation rules and the social behaviors being evaluated.

Boundary: A generated chain of thought is not evidence that the policy causally reasons that way. Successful trials and selected trajectory examples do not establish universal social safety.

Apply this lesson using the argument and evidence worksheet.

Inspection record and sources

Checked

2026-09-26

Reading status

Writing-focused manuscript scan + selected visuals inspected

Main text

Manuscript-wide writing scan of main PDF pp. 1–8: framing, section structure, method exposition, figure/table captions, evaluation organization and conclusions; not an exhaustive technical proof audit or reproduction.

Appendix / supplement

Separate supplement pp. 1–8 scanned: dataset construction, prompts/CoT example, architecture and training tables, reward ablations, extended open-loop tests and deployment images.

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.

• Official CVPR 2026 CVF proceedings PDF; retrieved 2026-09-26 · 11 pages
• Official CVPR 2026 separate supplementary PDF; retrieved 2026-09-26 · 8 pages