Skip to content
Computational Pathology Reading Roadmap

Writing study: Keep It Simple: Multi-Key Episodic Memory Retrieval for Ultra-Long Video Understanding

← ECCV 2026 oral writing corpus

Yeeun Choi, Youngbeom Yoo, Joon-Young Lee, Hyolim Kang, Seon Joo Kim · Official ECCV 2026 oral record

Figure, table and section ledger

Editorial contribution type: Efficiency, inference & 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
18 physical pages
Before references / appendix
≈ 15.15 page-equivalents
Main figures / tables
3 / 7
Attached appendix figures / tables
0 / 0
Attached appendix span
No boundary detected; not assumed zero
Before references / appendix: ≈ 84.17%References / intervening material: ≈ 15.83%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.

Source-bound correction: All top-level main headings and the reference boundary checked against source PDF text; source-bound positions use physical single-column y/page-height, not a two-column reading-order surrogate. Running page-number/author titles, affiliations, numbered contribution bullets, footnotes and bibliography continuations are excluded. This linked manuscript contains no appended supplementary section; separate supplement links were not scanned.

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 headingStarts on PDF pageEstimated spanShare of entire PDF
1 Introduction1≈ 2 pp.11.11%
2 Related Work3≈ 1.88 pp.10.44%
3 Method5≈ 3.08 pp.17.11%
4 Experiments8≈ 6.02 pp.33.44%
5 Qualitative Results14≈ 0.94 pp.5.22%
6 Conclusion15≈ 0.17 pp.0.94%
Acknowledgements15≈ 0.43 pp.2.39%
References16≈ 2.85 pp.15.83%
All detected figure and table locations

Main text

Attached appendix

    No recognized caption labels.

    Argument and organization

    1. 1 Introduction; 2 Related Work

      Starts with the infeasibility of sending hour-to-day videos directly to a model and questions whether complex offline memory structures are necessary.

    2. 3 Method: problem, memory keys and neighbor filtering

      Separates query-agnostic memory construction from query-dependent retrieval and reasoning, then tests key choice and temporal expansion independently.

    3. 4 Experiments; 5 Qualitative Results; 6 Conclusion

      The key-combination matrix sits next to matched-reproduction retrieval and memory-cost tables; captions identify reproduced baselines and solver choices.

    Problem, claim and contribution

    Starts with the infeasibility of sending hour-to-day videos directly to a model and questions whether complex offline memory structures are necessary.

    Abstract and Section 1, pp.1–3 · Main paper

    Argument and section order

    Separates query-agnostic memory construction from query-dependent retrieval and reasoning, then tests key choice and temporal expansion independently.

    Sections 3–4, pp.5–14 · Main paper

    Figures, captions and visual explanation

    The two-stage diagram uses example keys and an explicit answer-or-search-more branch, making offline storage and iterative online behavior visibly different.

    Figure 2, p.6 · Main paper

    Tables, comparisons and evidence

    The key-combination matrix sits next to matched-reproduction retrieval and memory-cost tables; captions identify reproduced baselines and solver choices.

    Tables 4–6, p.12 · Main paper

    Appendix and reproducibility

    Main-text pointers assign literal key-extraction and neighbor-filtering prompts to the appendix. Those prompts were not independently retrieved or inspected.

    Section 4.2, p.9 · Main paper

    Clarity and writing conventions

    Terms for memory keys, retrieval hit rate and QA accuracy remain distinct, avoiding a claim that better retrieval necessarily solves every reasoning question.

    Sections 3.3–3.4 and 4.4, pp.7–8,11–13 · Main paper

    What to reuse—and what not to infer

    Transferable lesson: For retrieval systems, pair an offline/online flowchart with separate indexing cost, hit rate and task accuracy, and disclose solver dependencies.

    Boundary: Prompting proprietary solvers and captioners is part of the experimental system. Token and call counts are not automatically equal monetary costs across models.

    Apply this lesson using the argument and evidence worksheet.

    Inspection record and sources

    Checked
    2026-10-02
    Reading status
    Writing-focused manuscript scan + selected visuals inspected
    Main text
    Writing-oriented scan across the manuscript main argument, physical PDF pp.1–15: framing and section roadmaps, method/argument explanations, figure and table captions, comparison protocols, conclusions and limitations. Selected full-page rasters were inspected. This is not exhaustive line-by-line technical reading, proof verification or experimental reproduction.
    Appendix / supplement
    Main-manuscript pointers to appendix/supplement and the reference boundary were checked. Separate supplementary PDFs were not retrieved or inspected; their contents are not included in inspected-page counts.
    Visually inspected PDF pages
    6, 12

    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.