PulseAugur
EN
LIVE 09:01:02

New method optimizes fair aggregation of noisy crowdsourced labels

Researchers have developed a new method for aggregating noisy labels from crowdsourced data, focusing on fairness and demographic parity. The study analyzes existing aggregation techniques like Majority Vote and Optimal Bayesian, providing theoretical guarantees on fairness gap convergence. They also introduce a generalized post-processing algorithm to enforce strict demographic parity constraints, demonstrating its effectiveness on synthetic and real datasets. AI

IMPACT Introduces theoretical guarantees and practical methods for fairer AI model training data.

RANK_REASON This is a research paper published on arXiv detailing a new method for aggregating noisy labels with fairness constraints. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

COVERAGE [1]

  1. arXiv cs.LG TIER_1 English(EN) · Gabriel Singer, Samuel Gruffaz, Olivier Vo Van, Nicolas Vayatis, Argyris Kalogeratos ·

    Optimal Fair Aggregation of Crowdsourced Noisy Labels using Demographic Parity Constraints

    arXiv:2601.23221v2 Announce Type: replace Abstract: As acquiring reliable ground-truth labels is usually costly, or infeasible, crowdsourcing and aggregation of noisy human annotations is the typical resort. Aggregating subjective labels, though, may amplify individual biases, pa…