Audit Algorithms
91 items from erwanlemerrer/awesome-audit-algorithms ★122
-
-
XRay: Enhancing the Web's Transparency with Differential Correlation usenix.org
(USENIX Security) Audits which user profile data were used for targeting a particular ad, recommendation, or price.
-
Stealing Machine Learning Models via Prediction APIs usenix.org
(Usenix Security) (Code) Aims at extracting machine learning models in use by remote services.
-
Data driven exploratory attacks on black box classifiers in adversarial domains sciencedirect.com
(Neurocomputing) Reverse engineers remote classifier models (e.g., for evading a CAPTCHA test).
-
Auditing fairness under unawareness through counterfactual reasoning sciencedirect.com
(Information Processing & Management) Shows how to unveil whether a black-box model, complying with the regulations, is still biased or not.
-
FairLens: Auditing black-box clinical decision support systems sciencedirect.com
(Information Processing & Management) Presents a pipeline to detect and explain potential fairness issues in Clinical DSS, by comparing different multi-label classification disparity measures.
-
Online Learning for Measuring Incentive Compatibility in Ad Auctions research.fb.com
(WWW) Measures the incentive compatible- (IC) mechanisms (regret) of black-box auction platforms.
-
Remote Explainability faces the bouncer problem rdcu.be
(Nature Machine Intelligence volume 2, pages529–539) (Code) Shows the impossibility (with one request) or the difficulty to spot lies on the explanations of a remote AI decision.
-
Identifying the Machine Learning Family from Black-Box Models rd.springer.com
(CAEPIA) Determines which kind of machine learning model is behind the returned predictions.
-
Look at the Variance! Efficient Black-box Explanations with Sobol-based Sensitivity Analysis proceedings.neurips.cc
(NeurIPS) Sobol indices provide an efficient way to capture higher-order interactions between image regions and their contributions to a (black box) neural network’s prediction through the lens of variance.
-
Black-Box Ripper: Copying black-box models using generative evolutionary algorithms proceedings.neurips.cc
(NeurIPS) Replicates the functionality of a black-box neural model, yet with no limit on the amount of queries (via a teacher/student scheme and an evolutionary search).
-
Active Fairness Auditing proceedings.mlr.press
(ICML) Studies of query-based auditing algorithms that can estimate the demographic parity of ML models in a query-efficient manner.
-
Bayesian Algorithm Execution: Estimating Computable Properties of Black-box Functions Using Mutual Information proceedings.mlr.press
(ICML) A budget constrained and Bayesian optimization procedure to extract properties out of a black-box algorithm.
-
Stealing the Decoding Algorithms of Language Models people.cs.umass.edu
(CCS) Steal the type and hyperparameters of the decoding algorithms of a LLM.
-
Opening Up the Black Box:Auditing Google's Top Stories Algorithm par.nsf.gov
(Flairs-32) Audit of the Google's Top stories panel that pro-vides insights into its algorithmic choices for selectingand ranking news publisher
-
Confidential-PROFITT: Confidential PROof of FaIr Training of Trees openreview.net
(ICLR) Proposes fair decision tree learning algorithms along with zero-knowledge proof protocols to obtain a proof of fairness on the audited server.
-
A zest of lime: towards architecture-independent model distances openreview.net
(ICLR) Measures the distance between two remote models using LIME.
-
GeoDA: a geometric framework for black-box adversarial attacks openaccess.thecvf.com
(CVPR) (Code) Crafts adversarial examples to fool models, in a pure blackbox setup (no gradients, inferred class only).
-
Auditing News Curation Systems:A Case Study Examining Algorithmic and Editorial Logic in Apple News ojs.aaai.org
(ICWSM) Audit study of Apple News as a sociotechnical news curation system (trending stories section).
-
Two-Face: Adversarial Audit of Commercial Face Recognition Systems ojs.aaai.org
(ICWSM) Performs an adversarial audit on multiple systems APIs and datasets, making a number of concerning observations.
-
Privacy Auditing with One (1) Training Run neurips.cc
(NeurIPS - best paper) A scheme for auditing differentially private machine learning systems with a single training run.
-
An Empirical Analysis of Algorithmic Pricing on Amazon Marketplace mislove.org
(WWW) (Code) Develops a methodology for detecting algorithmic pricing, and use it empirically to analyze their prevalence and behavior on Amazon Marketplace.
-
Modeling rabbit‑holes on YouTube link.springer.com
(SNAM) Models the trapping dynamics of users in rabbit holes in YouTube, and provides a measure of this enclosure.
-
Scaling up search engine audits: Practical insights for algorithm auditing journals.sagepub.com
(Journal of Information Science) (Code) Audits multiple search engines using simulated browsing behavior with virtual agents.
-
Query Strategies for Evading Convex-Inducing Classifiers jmlr.org
(JMLR) Evasion methods for convex classifiers. Considers evasion complexity.
-
Membership Inference Attacks Against Machine Learning Models ieeexplore.ieee.org
(Symposium on Security and Privacy) Given a machine learning model and a record, determine whether this record was used as part of the model's training dataset or not.
-
Algorithmic Transparency via Quantitative Input Influence: Theory and Experiments with Learning Systems ieeexplore.ieee.org
(Security and Privacy) Introduces measures that capture the degree of influence of inputs on outputs of the observed system.
-
Back in Black: Towards Formal, Black Box Analysis of Sanitizers and Filters ieeexplore.ieee.org
(Security and Privacy) Black-box analysis of sanitizers and filters.
-
Stealing Knowledge from Protected Deep Neural Networks Using Composite Unlabeled Data ieeexplore.ieee.org
(ICNN) Composite method which can be used to attack and extract the knowledge ofa black box model even if it completely conceals its softmaxoutput.
-
Under manipulations, are some AI models harder to audit? grodino.github.io
(SATML) Relates the difficulty of black-box audits to the capacity of the targeted models, using the Rademacher complexity.
- next page of items loading…
