Algorithmic Fairness Verification with Graphical Models
نویسندگان
چکیده
In recent years, machine learning (ML) algorithms have been deployed in safety-critical and high-stake decision-making, where the fairness of is paramount importance. Fairness ML centers on detecting bias towards certain demographic populations induced by an classifier proposes algorithmic solutions to mitigate with respect different definitions. To this end, several verifiers proposed that compute prediction classifier—essentially beyond a finite dataset—given probability distribution input features. context verifying linear classifiers, existing are limited accuracy due imprecise modeling correlations among features scalability restrictive formulations classifiers as SSAT/SMT formulas or sampling. paper, we propose efficient verifier, called FVGM, encodes Bayesian network. contrast verifiers, FVGM stochastic subset-sum based approach for classifiers. Experimentally, show leads accurate scalable assessment more diverse families fairness-enhancing algorithms, attacks, group/causal metrics than state-of-the-art verifiers. We also demonstrate facilitates computation influence functions stepping stone detect source subsets
منابع مشابه
Modular Verification of Qualitative Pathway Models with Fairness
Modular verification is a technique used to face the state explosion problem often encountered in the verification of properties of complex systems such as concurrent interactive systems. The modular approach is based on the observation that properties of interest often concern a rather small portion of the system. As a consequence, reduced models can be constructed which approximate the overal...
متن کاملLogical and Algorithmic Properties of Conditional Independence and Graphical Models
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. Institute of Mathematical Statistics is collaborating with...
متن کاملAlgorithmic Verification with Multiple and Nested Parameters
We consider parameterised verification problem, where parameters are sets and relations over these sets, typically used to denote sets of identities of replicated components and connections between the components. A specification and a system are given as (multiply) parameterised labelled transition systems, parameter values are encoded using first-order logic and correctness is understood as t...
متن کاملLearning with Graphical Models
Probabilistic graphical models are being used widely in artiicial intelligence, for instance, in diagnosis and expert systems, as a uniied qualitative and quantitative framework for representing and reasoning with probabilities and independencies. Their development and use spans several elds including artiicial intelligence, decision theory and statistics, and provides an important bridge betwe...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: Proceedings of the ... AAAI Conference on Artificial Intelligence
سال: 2022
ISSN: ['2159-5399', '2374-3468']
DOI: https://doi.org/10.1609/aaai.v36i9.21187