sensys-review-process
GitHub解析SenSys双截稿、自包含审稿机制，解读拒稿后需实质性修改并附回复方可重投的规则。分析评审对实测与能效的侧重，指导分类通知内容以决定接受、修改重投或转投其他会议。
触发场景
安装
npx skills add brycewang-stanford/Awesome-Journal-Skills --skill sensys-review-process -g -y
SKILL.md
Frontmatter
{
"name": "sensys-review-process",
"description": "Use when interpreting SenSys's two-deadline, self-contained review model — what a first-deadline outcome means, that a reject may return at the second deadline only with a substantive revision and Response to Reviewers, how systems reviewers weigh energy and deployment evidence, and how to classify a notification before drafting a response."
}
SenSys Review Process
SenSys does not run a single annual verdict. It runs two submission deadlines per cycle, each a self-contained review round, and the connective tissue between them is a resubmission rule: a paper rejected at one deadline may re-enter at the next only with a substantive revision and a "Response to Reviewers" that maps each concern to a specific change. Reading your notification correctly means knowing which move the process is offering you.
The outcome space
| Outcome | What it means | Your move |
|---|---|---|
| Accept | The paper is in the edition the deadline feeds | sensys-artifact-evaluation + sensys-camera-ready |
| Reject | Not accepted at this deadline | Decide whether to resubmit with revision at the next deadline, or route elsewhere |
| Resubmit with revision (next deadline) | The path the two-deadline model offers, not an automatic invitation | Treat the reviews as the required-changes list; see sensys-author-response |
Note that whether a given cycle also offers an explicit major-revision channel at notification (distinct from the first→second deadline resubmission) is 待核实 — read the current CFP; do not assume a rebuttal or revision stage that the edition does not run.
What SenSys reviewers weigh
SenSys reviewers are systems people who have built and deployed. They read for:
- Is it built and measured? Simulation-only or single-run results draw immediate skepticism.
- Is the energy real? Adjectives ("low-power") where a measurement belongs are flagged.
- Is the ground truth honest? Accuracy against an unstated or uncalibrated reference is a standard objection.
- Is the baseline fair? Cross-hardware or untuned comparisons undercut the headline.
- Does the design understand its space? Rejected alternatives and named limitations signal competence; their absence signals under-evaluation.
Knowing this lets you predict the review themes before they arrive and pre-empt them in the paper
(sensys-experiments, sensys-writing-style).
Classifying a notification before you respond
Read every review and sort each point:
[Blocking] a claim the reviewer does not believe without new measurement
[Fixable] a gap addressable by a bench run or a clearer figure
[Framing] a misread you can prevent with rewriting, no new data
[Wishlist] a nice-to-have outside the paper's scope
Then decide:
Accept → move to camera-ready + artifacts
Resubmit → is the blocking set closable by the next deadline given testbed time?
Route away → if the blocking set needs a redesign, a different venue may be faster
The decision to resubmit is a calendar-and-testbed decision as much as a scientific one: a
blocking reviewer objection that requires a new energy-harvesting deployment cannot be answered
in two weeks. sensys-workflow holds the backward schedule.
Confidentiality and conduct
Reviews are confidential; do not circulate them or identify reviewers, and do not attempt to de-anonymize. Concerns about a review's fairness or a conflict go to the chairs through the process, not into the response text. The exact conduct and appeal wording for the cycle is 待核实 — read the current pages.
Reviewer continuity across deadlines
Because a resubmission returns to the venue with its prior reviews attached, assume some
continuity of reviewer context between deadlines: the response-to-reviewers is read against
what was promised. This makes the response a contract (see sensys-author-response), not a
debate — every promised change must appear in the revised paper.
Output format
[Outcome] accept / reject / resubmit-with-revision — as the notification states
[Blocking] the points that must be answered with new measurement, listed
[Feasible] can the blocking set close by the next deadline given testbed time? Y/N
[Path] resubmit here / route elsewhere / proceed to camera-ready
[Open] any process ambiguity (major-revision channel? appeal?) marked 待核实
版本历史
- 9f86f09 当前 2026-07-19 17:41
