Hacktoberfest 2026: le issue che i maintainer hanno segnato per ottobre, aperte e adatte ai principianti. Sfoglia le issue Hacktoberfest

Explainers: Randomized Post-Processing, Fairness-Accuracy Pareto Frontiers and Choosing a Mitigation Strategy

Aperta
  • Qualcosa non va?
#1,102 0 commenti 0 reazioni 0 assegnatari Vedi su GitHub

I maintainer di solito rispondono entro 1 giorno

Nessuno ha ancora preso questa issue.

Valutazione

Difficoltà
5/5
Tempo stimato
Più di una settimana
Idoneità per principianti
25/100
Tipo di issue
Documentazione
Chiarezza
Specificata chiaramente
Stato di attività
Attiva
Stack tecnologico
json, markdown, python
Ambito
data, documentation, machine-learning

Direzione di ricerca

Read CONTRIBUTING.md and a close sibling explainer, then inspect results/ and the audit loading entry point faircode.benchmark.load_audit. Verify the cited papers and compute the requested real audit results for the three explainers. Done means updating the listed documentation and website entries, regenerating files with make build-explainers, and passing make lint.

Scritto dal modello di indicizzazione a partire dal testo della issue.

Descrizione

documentation help wanted

Mitigation Strategies introduces the benchmark's five strategies and Group-Specific Thresholds shows the simplest post-processing rule. Three questions remain for someone who has to pick one. Add all three in one pull request.

explainers/randomized-post-processing.md - What Is Randomized Post-Processing?

When two thresholds cannot reach an exact criterion, mix them: flip a weighted coin between two decisions for a group. Show it on a real audit with the equalized-odds construction (the mixing probabilities solve a small linear programme over the group ROC curves), compare it with deterministic group thresholds and with the benchmark's S4 row in results/, and cover the fairness objection to deciding by chance.

Key citation to verify: Hardt, Price, Srebro (2016), "Equality of Opportunity in Supervised Learning" (NeurIPS) - confirm.

explainers/fairness-accuracy-pareto-frontier.md - What Is a Fairness-Accuracy Pareto Frontier?

Instead of one model, the set of models that are not beaten on both accuracy and a fairness gap. Trace it on a real audit by sweeping a penalty or threshold, plot or tabulate the non-dominated points, and show where each of the benchmark's strategies lies against it. Cover why the frontier is a choice set, not an answer.

Key citation to verify: Martinez, Bertran, Sapiro (2020), "Minimax Pareto Fairness: A Multi Objective Perspective" (ICML) and Wei, Niethammer (2022), "The Fairness-Accuracy Pareto Front" (Statistical Analysis and Data Mining) - confirm.

explainers/choosing-a-mitigation-strategy.md - How Do You Choose Among Mitigation Strategies?

A decision guide built from the benchmark's own results: for each of the five strategies, what it needs (group labels at training, at prediction), what it costs in accuracy across the seven audits in results/, and what it leaves behind. Include a decision table and a computed summary (median accuracy cost and gap reduction per strategy) read straight from the CSV.

Key citation to verify: Cite the Fairlearn paper, Bird et al. (2020), "Fairlearn: A toolkit for assessing and improving fairness in AI" (Microsoft Research tech report) - confirm.

Related explainers: mitigation-strategies, group-specific-thresholds, reject-option-classification, fairness-accuracy-tradeoff, maxmin-fairness, equalized-odds.

See CONTRIBUTING.md: add explainers/<slug>.md plus a one-line entry in assets/explainers-data.json, then run make build-explainers; the build script generates the pages, sitemap, llms-full.txt and OG images, and llms.txt and the README table must list each explainer (scripts/check_explainer_count.py checks both). Match the structure of a close sibling explainer, use real code run against one of the repo's audits (faircode.benchmark.load_audit loads an audit exactly like the benchmark harness) and real, not invented, results, and if you quote a Fair Code result say whether it comes from results/ (current) or paper/results-frozen/ (the earlier reference snapshot), per CLAUDE.md. Verify every citation yourself before using it.

Done when: the 3 explainers, their README, CONTRIBUTING, ROADMAP, llms.txt and CHANGELOG entries, the website JSON entries and the regenerated generated files are merged, make lint passes, and each explainer's code is runnable from the repo root.

Lingua principale
HTML
Stelle
47
Fork
45
Merge medio
3h 12m
PR unite (30g)
77

Preparare l'ambiente

Come iniziare

  1. Leggi tutta la issue e poi la guida ai contributi del progetto.
  2. Commenta sulla issue per dire che te ne occupi tu — evita che due persone facciano lo stesso lavoro.
  3. Fai un fork del repository e lavora su un branch.
  4. Apri una pull request che faccia riferimento al numero della issue.

Altre issue di yakew7/Fair-Code

Tutte le issue di yakew7/Fair-Code

Issue simili

Altre issue su Data Engineering

Ricevi le nuove issue nella tua casella

Un breve riepilogo di issue GitHub adatte ai principianti.