hri-review-process

HRI Review Process

Understanding HRI's pipeline tells you where your effort actually moves the outcome. HRI runs a double-blind, two-phase review with a real rebuttal and a program-committee meeting — closer to a rigorous conference model than to a journal R&R, but with a discussion culture that gives the rebuttal genuine weight. Contribution-typed subcommittees mean your paper is judged by reviewers matched to its track. This skill models the pipeline; hri-author-response covers writing the rebuttal. Mechanics below are the HRI 2026 cycle (see resources/official-source-map.md).

The pipeline, stage by stage

1. Initial check (desk stage). Program-Committee Chairs, Track Chairs, and Area Chairs screen submissions for desk rejections (over length, anonymity violations, out of scope) and track changes (moving a paper to a better-fitting subcommittee). Clearing this stage is entirely in your control — see hri-submission.
2. Assignment. Each paper is assigned within its track/subcommittee to a 1AC (primary area chair) and a 2AC (secondary). The 1AC recruits three external reviewers — experts in the topic who are not on the program committee.
3. Reviewing. The three externals write reviews; the ACs read the paper and the reviews. So a typical paper has three external reviews plus two AC perspectives.
4. Rebuttal (phase two for authors). Authors receive the reviews and submit one rebuttal (HRI 2026: due ~12 Nov 2025), addressing misunderstandings, factual errors, and specific concerns.
5. Discussion. After the rebuttal, the 1AC leads an online discussion among the reviewers, who may revise scores in light of the rebuttal and each other's points.
6. Program-committee meeting. The PC meets to make final decisions, with the ACs championing or critiquing papers based on reviews, rebuttal, and discussion.
7. Decision (HRI 2026: December). Some editions may use light shepherding/conditional acceptance (待核实 per cycle). There is no in-cycle major-revision round — a reject is a reject for that edition.

What reviewers are told to value

HRI reviewer guidance steers toward impact over polish: reviewers are asked to favor slightly flawed but impactful work over perfectly executed but low-impact work, and to judge the paper as the contribution type of its track (a Design paper by design criteria, a User Studies paper by study rigor, and so on). The reviewer form asks reviewers to identify the paper's key findings and claimed contribution and to assess how well the evidence supports them. Write so those are effortless to state — an ambiguous contribution is a reviewing risk.

Where author leverage actually is

• Before submission (highest leverage): clear the desk check, pick the right track, make the contribution and its evidence unmistakable, and include an anonymized video. Most outcomes are set here.
• The rebuttal (real, bounded leverage): because reviewers discuss and can move scores after the rebuttal, a focused rebuttal that corrects a genuine misreading or a factual error routinely changes decisions. It is not a formality. See hri-author-response.
• You cannot influence: AC/reviewer assignment, the discussion you do not see, or the PC-meeting debate. Do not waste the rebuttal arguing with tone; fix what a reviewer can act on.

Calibrating your expectations

• HRI is highly selective — recent editions accepted roughly a quarter of full-paper submissions (HRI 2025 reported ~100/400 ≈ 25%; HRI 2024 ~24.7%; treat exact figures as reported, 待核实 to the official program). A strong paper still faces a hard bar.
• Three externals + two ACs means disagreement is normal; a split (e.g., two positive, one negative) is common and is exactly where the rebuttal and discussion decide the paper.
• Best-paper awards are given per track, so excellence is judged within your contribution type, not across the whole conference.

Reading your reviews

• Separate fatal concerns (a design confound, an unsupported central claim, an ethics gap) from fixable ones (a missing citation, a clarification, an added analysis you can describe).
• Identify each reviewer's decisive objection — the one thing that, if resolved, flips them — and aim the rebuttal there.
• Note where reviewers disagree; the rebuttal can help the discussion resolve the split in your favor by giving the sympathetic reviewer ammunition.

Reverify each cycle

• Reviewer/AC counts (3 externals + 1AC/2AC is the recent norm) and the desk-check ownership.
• Whether the edition uses shepherding/conditional acceptance.
• Rebuttal timing, length, and whether referencing new results is allowed (待核实).

Output format

[Stage] where the paper is: desk-check / reviewing / rebuttal / discussion / PC-meeting / decided
[Reviewer picture] 3 externals + 1AC/2AC · scores/split summarized
[Decisive objections] per reviewer, the one thing that flips them
[Leverage] what the rebuttal can realistically move here
[Track] judged as which contribution type · award eligibility per track
[Next action] <ordered>