A Methodology for Non-Functional Property Evaluation of Machine Learning Models

Marco Anisetti, Claudio A. Ardagna, Ernesto Damiani, Paolo G. Panero

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

6 Scopus citations

Abstract

The pervasive diffusion of Machine Learning (ML) in many critical domains and application scenarios has revolutionized implementation and working of modern IT systems. The behavior of modern systems often depends on the behavior of ML models, which are treated as black boxes, thus making automated decisions based on inference unpredictable. In this context, there is an increasing need of verifying the non-functional properties of ML models, such as, fairness and privacy, to the aim of providing certified ML-based applications and services. In this paper, we propose a methodology based on Multi-Armed Bandit for evaluating non-functional properties of ML models. Our methodology adopts Thompson sampling, Monte Carlo Simulation, and Value Remaining. An experimental evaluation in a real-world scenario is presented to prove the applicability of our approach in evaluating the fairness of different ML models.

Original languageBritish English
Title of host publicationProceedings of the 12th International Conference on Management of Digital EcoSystems, MEDES 2020
Pages38-45
Number of pages8
ISBN (Electronic)9781450381154
DOIs
StatePublished - 2 Nov 2020
Event12th International Conference on Management of Digital EcoSystems, MEDES 2020 - Virtual, Online, United Arab Emirates
Duration: 2 Nov 20204 Nov 2020

Publication series

NameProceedings of the 12th International Conference on Management of Digital EcoSystems, MEDES 2020

Conference

Conference12th International Conference on Management of Digital EcoSystems, MEDES 2020
Country/TerritoryUnited Arab Emirates
CityVirtual, Online
Period2/11/204/11/20

Keywords

  • Machine Learning Assurance
  • Multiarmed bandit
  • Non-functional properties

Fingerprint

Dive into the research topics of 'A Methodology for Non-Functional Property Evaluation of Machine Learning Models'. Together they form a unique fingerprint.

Cite this