Abstract:Professional networks provide invaluable entree to opportunity through referrals and introductions. A rich literature shows they also serve to entrench and even exacerbate a status quo of privilege and disadvantage. Hiring platforms, equipped with the ability to nudge link formation, provide a tantalizing opening for beneficial structural change. We anticipate that key to this prospect will be the ability to estimate the likelihood of edge formation in an evolving graph. Outcome-indistinguishable prediction algorithms ensure that the modeled world is indistinguishable from the real world by a family of statistical tests. Omnipredictors ensure that predictions can be post-processed to yield loss minimization competitive with respect to a benchmark class of predictors for many losses simultaneously, with appropriate post- processing. We begin by observing that, by combining a slightly modified form of the online K29 star algorithm of Vovk (2007) with basic facts from the theory of reproducing kernel Hilbert spaces, one can derive simple and efficient online algorithms satisfying outcome indistinguishability and omniprediction, with guarantees that improve upon, or are complementary to, those currently known. This is of independent interest. We apply these techniques to evolving graphs, obtaining online outcome-indistinguishable omnipredictors for rich -- possibly infinite -- sets of distinguishers that capture properties of pairs of nodes, and their neighborhoods. This yields, inter alia, multicalibrated predictions of edge formation with respect to pairs of demographic groups, and the ability to simultaneously optimize loss as measured by a variety of social welfare functions.

Ahead of the Count: An Algorithm for Probabilistic Prediction of Instant Runoff (IRV) Elections

The probability of casting a pivotal vote in an Instant Runoff Voting election

Ballot-Polling Audits of Instant-Runoff Voting Elections with a Dirichlet-Tree Model

Optimal Strategies in Ranked Choice Voting

The Moderating Effect of Instant Runoff Voting

Efficient Weighting Schemes for Auditing Instant-Runoff Voting Elections

Ballot Length in Instant Runoff Voting

Auditing Ranked Voting Elections with Dirichlet-Tree Models: First Steps

Adaptively Weighted Audits of Instant-Runoff Voting Elections: AWAIRE

RAIRE: Risk-Limiting Audits for IRV Elections

Forecasting elections from partial information using a Bayesian model for a multinomial sequence of data

Improving the Computational Efficiency of Adaptive Audits of IRV Elections

New Algorithms and Applications for Risk-Limiting Audits

Quickly Determining Who Won an Election

Generalizing Instant Runoff Voting to Allow Indifferences

A Comparison of Sequential Ranked-Choice Voting and Single Transferable Vote

Finding the Fairest Voting System using Likelihood Analysis

Risk-Limiting Audits for Condorcet Elections

From Fairness to Infinity: Outcome-Indistinguishable (Omni)Prediction in Evolving Graphs

Evolving Accuracy: A Genetic Algorithm to Improve Election Night Forecasts

Distributed Monitoring of Election Winners