riskcal

Documentation

  • API Reference
    • riskcal.analysis
      • PLD (Privacy Loss Distribution)
        • get_beta_from_pld()
        • get_advantage_from_pld()
        • get_bayes_risk_from_pld()
      • GDP (Gaussian Differential Privacy)
        • get_beta_from_gdp()
        • get_advantage_from_gdp()
        • get_bayes_risk_from_gdp()
      • ADP (Approximate Differential Privacy)
        • get_beta_from_adp()
        • get_advantage_from_adp()
        • get_epsilon_from_err_rates()
      • RDP (Renyi Differential Privacy)
        • get_beta_from_rdp()
      • zCDP (Zero-Concentrated Differential Privacy)
        • get_beta_from_zcdp()
        • get_advantage_from_zcdp()
      • Internal utilities
        • pld_to_plrvs()
        • PLRVs
    • riskcal.calibration
      • Core calibration interface
        • calibrate_parameter()
        • CalibrationTarget
        • CalibrationConfig
        • CalibrationResult
        • PrivacyMetrics
        • PrivacyEvaluator
      • DP-SGD calibration
        • find_noise_multiplier_for_advantage_dpsgd()
        • find_noise_multiplier_for_err_rates_dpsgd()
        • get_advantage_for_dpsgd()
        • get_beta_for_dpsgd()
        • create_dpsgd_evaluator()
        • create_dpsgd_epsilon_evaluator()
      • Blackbox calibration
        • find_noise_multiplier_for_epsilon_delta()
        • find_noise_multiplier_for_advantage_blackbox()
        • find_noise_multiplier_for_err_rates_blackbox()
        • create_accountant_evaluator()
    • riskcal.accountants
      • CTDAccountant
        • CTDAccountant

Links

  • GitHub Repository
  • PyPI Package
  • Paper (NeurIPS 2024)
  • Paper (NeurIPS 2025)
riskcal
  • Search

© Copyright 2025, Bogdan Kulynych, Juan Felipe Gomez.

Built with Sphinx using a theme provided by Read the Docs.