Olivier Binette
Statistician working on entity resolution and its evaluation; lead author of
er-evaluation-framework and maintainer of the er-evaluation package it releases.
The relevant detail for this wiki is that his contribution is measurement of the measurement. His own earlier work (Binette et al., 2022) is what the 2024 paper cites for the finding that naive pairwise precision on benchmark data is biased upward, and the framework paper is the constructive answer to it — an estimator whose bias is itself reported (under 0.4%). That is the discipline the corpus’s entity-resolution thread has been missing: not a better merger, a trustworthy way to score one.