ExpProof : Operationalizing Explanations for Confidential Models with ZKPs
Not on the map. Why?
Abstract (arXiv)
In principle, explanations are intended as a way to increase trust in machine learning models and are often obligated by regulations. However, many circumstances where these are demanded are adversarial in nature, meaning the involved parties have misaligned interests and are incentivized to manipulate explanations for their purpose. As a result, explainability methods fail to be operational in such settings despite the demand citebordt2022post. In this paper, we take a step towards operationalizing explanations in adversarial scenarios with Zero-Knowledge Proofs (ZKPs), a cryptographic primitive. Specifically we explore ZKP-amenable versions of the popular explainability algorithm LIME and evaluate their performance on Neural Networks and Random Forests. Our code is publicly available at https://github.com/emlaufer/ExpProof.
Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Builds on 28 of its 68 references that are in the map. First posted on arXiv 6 Feb 2025.
Builds on
28 of 68 references.
References · 28
- 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,305
- 2016"Why Should I Trust You?": Explaining the Predictions of Any Classifiercited by 5,070
- 2017Counterfactual Explanations without Opening the Black Box: Automated Decisions and the GDPRcited by 1,089
- 2019Fooling LIME and SHAP: Adversarial Attacks on Post hoc Explanation Methodscited by 443
- 2021Human-Centered Explainable AI (XAI): From Algorithms to User Experiencescited by 123
- 2021What Do We Want From Explainable Artificial Intelligence (XAI)? -- A Stakeholder Perspective on XAI and a Conceptual Model Guiding Interdisciplinary XAI Researchcited by 119
- 2020A survey of algorithmic recourse: definitions, formulations, solutions, and prospectscited by 118
- 2021Counterfactual Explanations Can Be Manipulatedcited by 113
- 2020Explaining the Explainer: A First Theoretical Analysis of LIMEcited by 89
- 2019Fairwashing: the risk of rationalizationcited by 81
- 2022Framework for Evaluating Faithfulness of Local Explanationscited by 55
- 2022Post-Hoc Explanations Fail to Achieve their Purpose in Adversarial Contextscited by 53
16 more
- 2024zkLLM: Zero Knowledge Proofs for Large Language Modelscited by 51
- 2017Inverse Classification for Comparison-based Interpretability in Machine Learningcited by 50
- 2018Defining Locality for Surrogates in Post-hoc Interpretablitycited by 33
- 2022Scaling up Trustless DNN Inference with Zero-Knowledge Proofscited by 31
- 2022Active Fairness Auditingcited by 27
- 2019Provable Certificates for Adversarial Examples: Fitting a Ball in the Union of Polytopescited by 25
- 2020Looking Deeper into Tabular LIMEcited by 20
- 2022pvCNN: Privacy-Preserving and Verifiable Convolutional Neural Network Testingcited by 20
- 2021On the Relation of Trust and Explainability: Why to Engineer for Trustworthinesscited by 16
- 2021Fairness, Integrity, and Privacy in a Scalable Blockchain-based Federated Learning Systemcited by 16
- 2024Trustless Audits without Revealing Data or Modelscited by 15
- 2024FairProof : Confidential and Certifiable Fairness for Neural Networkscited by 13
- 2024Auditing Local Explanations is Hardcited by 8
- 2023Verifiable Fairness: Privacy-preserving Computation of Fairness for Machine Learning Systemscited by 7
- 2022XAudit : A Theoretical Look at Auditing with Explanationscited by 5
- 2024Influence-based Attributions can be Manipulatedcited by 2
Built on by
Citing papers per year: 2026: 2.
2 papers.