AI Paper Atlas

Research map / Learning theory and algorithms / Prediction and decision losses

Designing Rules to Pick a Rule: Aggregation by Consistency

Ratip Emin Berker, Ben Armstrong, Vincent Conitzer, Nihar B Shah

ICLR 2026 · Learning theory and algorithms › Prediction and decision losses · approach: Ranking models from comparisons

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
pick the rule minimizing disagreement across random splits (AbC) on data from Mallows or Plackett-LuceAbC · synthetic partial rankings (100 voters, 100 items)split disagreement and distance to ground truthImproved
aggregate review scores by Max instead of Arithmetic Mean (ablation)telescope proposal review scores · observatory telescope time proposalsnormalized Kendall-Tau split distance: 0.364 -> 0.409Hurt
aggregate review scores by Min instead of Arithmetic Mean (ablation)telescope proposal review scores · observatory telescope time proposalsnormalized Kendall-Tau split distance: 0.364 -> 0.444Hurt
use Trimmed Mean instead of Arithmetic Mean (ablation)telescope proposal review scores · observatory telescope time proposalsnormalized Kendall-Tau split distance: 0.364 -> 0.371No clear effect
use Trimmed Borda instead of Borda count (ablation)ALMA proposal rankings · ALMA Cycle 10 and 11 reviewssplit disagreementHurt
use Kemeny rule instead of Borda or PL MLE (ablation)ALMA proposal rankings · ALMA Cycle 10 and 11 reviewssplit disagreementHurt
use rules using more positions (Borda, optimized positional scores) instead of Plurality or Veto (ablation)ALMA proposal rankings · ALMA Cycle 10 and 11 reviewssplit disagreementImproved
use IRV as used in practice versus other rules (ablation)US municipal elections (25 Preflib elections) · Preflib IRV electionssplit disagreementHurt
update F1 scoring rule in 2003 and in 2010 (ablation)Formula One scoring · F1 seasons 1991-2018split disagreementImproved
optimize the positional scoring vector instead of using F1 rules (ablation)F1 race rankings · F1 seasons 1991-2018split disagreementImproved
pick welfare-maximizing rule instead of consistency-maximizing rulerule picking rules · theoryplurality-shuffling consistencyHurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy