Model Reconstruction from Model Explanations
Open accessThe DOI matches its Crossref record; 10 of the 10 references it cites were found in a registry or an index.
Abstract
We show through theory and experiment that gradient-based explanations of a model quickly reveal the model itself. Our results speak to a tension between the desire to keep a proprietary model secret and the ability to offer model explanations.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474
Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps
Simonyan, K., Vedaldi, A., & Zisserman, A. (2013) · arXiv · doi:10.48550/arxiv.1312.6034 · Cited by 5,370 · cited 1×
“The most basic saliency map is to compute the gradient of the model with respect to a chosen input [2, 13].”
BackgroundMembership Inference Attacks Against Machine Learning Models
Shokri, R., Stronati, M., Song, C., & Shmatikov, V. (2017) · 2017 IEEE Symposium on Security and Privacy (SP) · doi:10.1109/sp.2017.41 · Cited by 4,754 · cited 1×
“Yet another reason is that the model might leak sensitive information about the data it was trained on [5, 12].”
BackgroundAxiomatic attribution for deep networks
Sundararajan, M., Taly, A., & Yan, Q. (2017) · arXiv · doi:10.48550/arxiv.1703.01365 · Cited by 4,087
Practical Black-Box Attacks against Machine Learning
Papernot, N., McDaniel, P., Goodfellow, I., Jha, S., Celik, Z. B., & Swami, A. (2017) · Proceedings of the 2017 ACM on Asia Conference on Computer and Communications Security · doi:10.1145/3052973.3053009 · Cited by 3,495 · cited 1×
“In order to compare to reconstructing with membership queries alone, our method for learning with gradients is a modification of a simple heuristic used to reconstruct models from membership queries: training a new classifier f̂ to match the outputs of f [10, 19].”
MethodStriving for Simplicity: The All Convolutional Net
Springenberg, J. T., Dosovitskiy, A., Brox, T., & Riedmiller, M. (2014) · arXiv · doi:10.48550/arxiv.1412.6806 · Cited by 3,447
SmoothGrad: removing noise by adding noise
Smilkov, D., Thorat, N., Kim, B., Viégas, F., & Wattenberg, M. (2017) · arXiv · doi:10.48550/arxiv.1706.03825 · Cited by 1,467 · cited 1×
“and gradients processed with SmoothGrad, a saliency map denoising technique [15].”
MethodStealing Machine Learning Models via Prediction APIs
Tramèr, F., Zhang, F., Juels, A., Reiter, M. K., & Ristenpart, T. (2016) · arXiv · doi:10.48550/arxiv.1609.02943 · Cited by 796
The Secret Sharer: Evaluating and Testing Unintended Memorization in Neural Networks
Carlini, N., Liu, C., Erlingsson, Ú., Kos, J., & Song, D. (2018) · arXiv · doi:10.48550/arxiv.1802.08232 · Cited by 629 · cited 1×
“Yet another reason is that the model might leak sensitive information about the data it was trained on [5, 12].”
BackgroundHow to Explain Individual Classification Decisions
(2010) · Journal of Machine Learning Research · doi:10.5555/1756006.1859912 · Cited by 54 · cited 1×
“The most basic saliency map is to compute the gradient of the model with respect to a chosen input [2, 13].”
Background
Cited by
Explainable machine learning in deployment
Bhatt, U., Xiang, A., Sharma, S., Weller, A., Taly, A., Jia, Y., Ghosh, J., Puri, R., Moura, J. M. F., & Eckersley, P. (2020) · Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency · doi:10.1145/3351095.3375624 · Cited by 633
Actionable Recourse in Linear Classification
Ustun, B., Spangher, A., & Liu, Y. (2019) · Proceedings of the Conference on Fairness, Accountability, and Transparency · doi:10.1145/3287560.3287566 · Cited by 501
Hsieh, C.-Y., Li, C.-L., Yeh, C.-K., Nakhost, H., Fujii, Y., Ratner, A., Krishna, R., Lee, C.-Y., & Pfister, T. (2023) · Findings of the Association for Computational Linguistics: ACL 2023 · doi:10.18653/v1/2023.findings-acl.507 · Cited by 316
A Survey of Algorithmic Recourse: Contrastive Explanations and Consequential Recommendations
Karimi, A.-H., Barthe, G., Schölkopf, B., & Valera, I. (2023) · ACM Computing Surveys · doi:10.1145/3527848 · Cited by 301
A Survey of Privacy Attacks in Machine Learning
Rigaki, M., & Garcia, S. (2024) · ACM Computing Surveys · doi:10.1145/3624010 · Cited by 285
“Some, mostly theoretical, work has demonstrated the ability to perform direct model extraction beyond linear models [50, 86].”
BackgroundMethodInfluential
Metadata
- Authors
- Milli, S., Schmidt, L., Dragan, A. D., & Hardt, M.
- Venue
- Proceedings of the Conference on Fairness, Accountability, and Transparency
- Publisher
- ACM
- Published
- 2019-01-29
- Pages
- 1-9
- OpenAlex
- W2884148825
- Cited by
- 141
- References
- 10
- FWCI
- 7.21
- Type
- conference paper
- Language
- en
- Open access
- green
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.