How to use TaxoReview

Where to start, then a tutorial with screenshots for each page of a venue.
7

Diagnose & recruit

Menu: Recruitment → Diagnose & recruit Open it in the demonstration →

For a venue that has a pool. The diagnosis finds the papers and the topics the pool cannot serve well, and proposes the qualified experts outside it who could: for all the gaps at once, for one paper, or for one topic.

The diagnosis of the SIGIR 2026 reviewers: 138 ill papers, 7 under-covered topics, and 44 experts who would close every gap.
The diagnosis of the SIGIR 2026 reviewers: 138 ill papers, 7 under-covered topics, and 44 experts who would close every gap.
  1. 1 Pick the role: reviewers and area chairs are diagnosed separately, each with its own bar and loads.
  2. 2 People per paper and papers per person come from the venue; the bar in force is shown next to them.
  3. 3 Run the diagnosis. It takes under a minute; run it again after the pool or the assignment changes.
  4. 4 The summary. A paper is ill when the pool cannot give it enough strong people while it serves the other papers too, has no specialist for it, or the assignment left it short or weakly matched. Strong means in the top 1% of the database for that paper.
  5. 5 Close the gaps: the shortest list of experts that fills the missing places on all ill papers, each with the papers they would take.
  6. 6 Shortlist them for invitation, or add them to the pool; invite a few more than you need.
Recruiting for one paper.
Recruiting for one paper.
  1. 1 Severity orders the ill papers, worst first.
  2. 2 The reasons a paper is ill.
  3. 3 Strong in pool against Can get. The first counts everyone in the pool who is a strong match for the paper, however busy. The second is how many of them the paper can actually have, of the number it needs, when every paper draws on the same people and nobody takes more than their load. A paper with strong people in the pool can still get none: other papers need them just as much.
  4. 4 Recruit for this paper opens the closest qualified experts who have no authorship or recent co-authorship with its authors.
  5. 5 Tick and shortlist (or add to the pool).
Recruiting for one topic.
Recruiting for one topic.
  1. 1 Under-covered topics: groups of similar submissions, named from LLM keywords, where at least 15% of the strong places cannot be filled.
  2. 2 Coverage of the topic's places, with the places missing and the ill papers in it.
  3. 3 Recruit for this topic opens the experts closest to the topic as a whole. conflict with some papers marks somebody who is an author or co-author on part of the topic: they can take its other papers.
  4. 4 Add the ticked people, then run the matching again so that they are assigned.
← Taxonomy & gaps Reviewers in the wild →