mlsanityData to 7 Oct 2026

Influence-based Attributions can be Manipulated

arXiv 2409.05208 ↗First posted 8 Sep 2024cs.LG cs.AIcited by 2

Not on the map. Why?

Abstract (arXiv)

Influence Functions are a standard tool for attributing predictions to training data in a principled manner and are widely used in applications such as data valuation and fairness. In this work, we present realistic incentives to manipulate influence-based attributions and investigate whether these attributions can be textitsystematically tampered by an adversary. We show that this is indeed possible for logistic regression models trained on ResNet feature embeddings and standard tabular fairness datasets and provide efficient attacks with backward-friendly implementations. Our work raises questions on the reliability of influence-based attributions in adversarial circumstances. Code is available at : urlhttps://github.com/infinite-pursuits/influence-based-attributions-can-be-manipulated

Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Builds on 30 of its 47 references that are in the map. First posted on arXiv 8 Sep 2024.

Builds on

30 of 47 references.

References · 30

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2017Understanding Black-box Predictions via Influence Functionscited by 1,714
  3. 2018Towards Robust Interpretability with Self-Explaining Neural Networkscited by 505
  4. 2017Interpretation of Neural Networks is Fragilecited by 443
  5. 2019Fooling LIME and SHAP: Adversarial Attacks on Post hoc Explanation Methodscited by 440
  6. 2017The (Un)reliability of saliency methodscited by 331
  7. 2019Towards Efficient Data Valuation Based on the Shapley Valuecited by 279
  8. 2020Influence Functions in Deep Learning Are Fragilecited by 247
  9. 2019Explanations can be manipulated and geometry is to blamecited by 215
  10. 2022If Influence Functions are the Answer, Then What is the Question?cited by 142
  11. 2021Scaling Up Influence Functionscited by 130
  12. 2023DataInf: Efficiently Estimating Data Influence in LoRA-tuned LLMs and Diffusion Modelscited by 121
18 more
  1. 2020FastIF: Scalable Influence Functions for Efficient Model Interpretation and Debuggingcited by 120
  2. 2019Fooling Neural Network Interpretations via Adversarial Model Manipulationcited by 112
  3. 2019Learning to Deceive with Attention-Based Explanationscited by 111
  4. 2018Interpretable Deep Learning under Firecited by 74
  5. 2020Fairwashing Explanations with Off-Manifold Detergentcited by 71
  6. 2023Fast Model Debias with Machine Unlearningcited by 70
  7. 2022PUMA: Performance Unchanged Model Augmentation for Training Data Removalcited by 56
  8. 2022Post-Hoc Explanations Fail to Achieve their Purpose in Adversarial Contextscited by 53
  9. 2024Amazing Things Come From Having Many Good Modelscited by 52
  10. 2021Interactive Label Cleaning with Example-based Explanationscited by 48
  11. 2022Achieving Fairness at No Utility Cost via Data Reweighing with Influencecited by 38
  12. 2019Less Is Better: Unweighted Data Subsampling via Influence Functioncited by 30
  13. 2022Fair Infinitesimal Jackknife: Mitigating the Influence of Biased Training Data Points Without Refittingcited by 26
  14. 2023Revisiting the Fragility of Influence Functionscited by 20
  15. 2023zkDL: Efficient Zero-Knowledge Proofs of Deep Learning Trainingcited by 20
  16. 2022FairIF: Boosting Fairness in Deep Learning via Influence Functions with Validation Set Sensitive Attributescited by 12
  17. 2019Rewarding High-Quality Data via Influence Functionscited by 11
  18. 2022How Biased are Your Features?: Computing Fairness Influence Functions with Global Sensitivity Analysiscited by 7

Built on by

Cited by year

Citing papers per year: 2025: 2.

2 papers.

  1. 2025ExpProof : Operationalizing Explanations for Confidential Models with ZKPscited by 1
  2. 2025Efficiently Attacking Memorization Scorescited by 0