The DOI matches its Crossref record; 30 of the 32 references it cites were found in a registry or an index.
Eternal Sunshine of the Spotless Net: Selective Forgetting in Deep Networks
Open accessAbstract
We explore the problem of selectively forgetting a particular subset of the data used for training a deep neural network. While the effects of the data to be forgotten can be hidden from the output of the network, insights may still be gleaned by probing deep into its weights. We propose a method for "scrubbing" the weights clean of information about a particular set of training data. The method does not require retraining from scratch, nor access to the data originally used for training. Instead, the weights are modified so that any probing function of the weights is indistinguishable from the same function applied to the weights of a network trained without the data to be forgotten. This condition is a generalized and weaker form of Differential Privacy. Exploiting ideas related to the stability of stochastic gradient descent, we introduce an upper-bound on the amount of information remaining in the weights, which can be estimated efficiently even for deep neural networks.
Overview
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Second, most of the above schemes focus on relatively simple machine learning problems, such as linear logistic regression, or complex models with special assumptions [40, 41, 43, 51].”
From Machine Unlearning: A Survey (2024) · cites this paper as [40]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Significantly, model manipulation methods usually give a verifiability guarantee using theory-based and information bound-based methods [40, 41, 43].”
From Machine Unlearning: A Survey (2024) · cites this paper as [40]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“The higher the degree of similarity, the better the effect of the unlearning process.”
From Machine Unlearning: A Survey (2024) · cites this paper as [40]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Unlearning based on model shifting usually removes the impact of the unlearning dataset by calculating the influence on the model [40, 43].”
From Machine Unlearning: A Survey (2024) · cites this paper as [40]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“The information bound metric relies on KL-divergence to measure the information remaining about the unlearning dataset within the model after the unlearning process.”
From Machine Unlearning: A Survey (2024) · cites this paper as [40]
The cited source says what the claim attributes to it.
In the citing paper
“For the data that need to be unlearned, the accuracy should ideally be the same as a model trained without seeing D u [40].”
From Machine Unlearning: A Survey (2024) · cites this paper as [40]
The cited source says what the claim attributes to it.
In the citing paper
“Golatkar et al. [40, 43] devised a new metric for verifying the effectiveness of unlearning schemes, where they measured the upper bound of the residual information about samples that need to be unlearned.”
From Machine Unlearning: A Survey (2024) · cites this paper as [40]
The cited source says what the claim attributes to it.
In the citing paper
“Golatkar et al. [40, 43, 52] verify the effectiveness of their unlearning schemes based on accuracy and relearning time.”
From Machine Unlearning: A Survey (2024) · cites this paper as [40]
The cited source says what the claim attributes to it.
In the citing paper
“Then they evaluated the similarity of those distributions against the confidence of a trained model that has never seen the unlearning dataset.”
From Machine Unlearning: A Survey (2024) · cites this paper as [40]
The cited source says what the claim attributes to it.
In the citing paper
“They also developed two new verification metrics: model confidence and information bound [40].”
From Machine Unlearning: A Survey (2024) · cites this paper as [40]
The cited source says what the claim attributes to it.
In the citing paper
“Model confidence is formulated by measuring the distribution of the entropy of the output predictions on the remaining dataset, the unlearning dataset, and the test dataset.”
From Machine Unlearning: A Survey (2024) · cites this paper as [40]
The check wasn’t confident either way, so it gives no verdict — never disputed.
In the citing paper
“This is a similar metric to model confidence [40].”
From Machine Unlearning: A Survey (2024) · cites this paper as [40]
Experimental — accuracy not yet measured
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
Learning Multiple Layers of Features from Tiny Images
Krizhevsky, A. (2024) · doi:10.57702/zp44cu3g · Cited by 23,889
Overcoming catastrophic forgetting in neural networks
Kirkpatrick, J., Pascanu, R., Rabinowitz, N., Veness, J., Desjardins, G., Rusu, A. A., Milan, K., Quan, J., Ramalho, T., Grabska-Barwinska, A., et al. (2017) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1611835114 · Cited by 7,899
Deep Learning with Differential Privacy
Abadi, M., Chu, A., Goodfellow, I., McMahan, H. B., Mironov, I., Talwar, K., & Zhang, L. (2016) · Proceedings of the 2016 ACM SIGSAC Conference on Computer and Communications Security · doi:10.1145/2976749.2978318 · Cited by 6,578
The Algorithmic Foundations of Differential Privacy
Dwork, C., & Roth, A. (2014) · Foundations and Trends® in Theoretical Computer Science · doi:10.1561/0400000042 · Cited by 5,557
The Algorithmic Foundations of Differential Privacy
Dwork, C., & Roth, A. (2013) · doi:10.1561/9781601988195 · Cited by 3,856
Striving 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
(1989) · Cited by 3,214
VGGFace2: A Dataset for Recognising Faces across Pose and Age
Cao, Q., Shen, L., Xie, W., Parkhi, O. M., & Zisserman, A. (2018) · 2018 13th IEEE International Conference on Automatic Face & Gesture Recognition (FG 2018) · doi:10.1109/fg.2018.00020 · Cited by 3,032
Model Inversion Attacks that Exploit Confidence Information and Basic Countermeasures
Fredrikson, M., Jha, S., & Ristenpart, T. (2015) · Proceedings of the 22nd ACM SIGSAC Conference on Computer and Communications Security · doi:10.1145/2810103.2813677 · Cited by 2,961
Understanding deep learning requires rethinking generalization
Zhang, C., Bengio, S., Hardt, M., Recht, B., & Vinyals, O. (2016) · arXiv · doi:10.48550/arxiv.1611.03530 · Cited by 2,838
Deep Models Under the GAN: Information Leakage from Collaborative Deep Learning
Hitaj, B., Ateniese, G., & Perez-Cruz, F. (2017) · Proceedings of the 2017 ACM SIGSAC Conference on Computer and Communications Security · doi:10.1145/3133956.3134012 · Cited by 1,377
A Closer Look at Memorization in Deep Networks
Arpit, D., Jastrzębski, S., Ballas, N., Krueger, D., Bengio, E., Kanwal, M. S., Maharaj, T., Fischer, A., Courville, A., Bengio, Y., et al. (2017) · arXiv · doi:10.48550/arxiv.1706.05394 · Cited by 1,307
New Insights and Perspectives on the Natural Gradient MethodPreprint
Martens, J. (2014) · Journal of machine learning research · doi:10.48550/arxiv.1412.1193 · Cited by 884
Variational Dropout and the Local Reparameterization TrickPreprint
Kingma, D. P., Salimans, T., & Welling, M. (2015) · arXiv (Cornell University) · doi:10.48550/arxiv.1506.02557 · Cited by 822
Bourtoule, L., Chandrasekaran, V., Choquette-Choo, C. A., Jia, H., Travers, A., Zhang, B., Lie, D., & Papernot, N. (2021) · 2021 IEEE Symposium on Security and Privacy (SP) · doi:10.1109/sp40001.2021.00019 · Cited by 810
Towards Making Systems Forget with Machine Unlearning
Cao, Y., & Yang, J. (2015) · 2015 IEEE Symposium on Security and Privacy · doi:10.1109/sp.2015.35 · Cited by 622
Dziugaite, G. K., & Roy, D. M. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1703.11008 · Cited by 599
Machine Learning Models that Remember Too Much
Song, C., Ristenpart, T., & Shmatikov, V. (2017) · Proceedings of the 2017 ACM SIGSAC Conference on Computer and Communications Security · doi:10.1145/3133956.3134077 · Cited by 526
Optimizing Neural Networks with Kronecker-factored Approximate CurvaturePreprint
Martens, J., & Grosse, R. (2015) · arXiv (Cornell University) · doi:10.48550/arxiv.1503.05671 · Cited by 476
LOGAN: Membership Inference Attacks Against Generative Models
Hayes, J., Melis, L., Danezis, G., & De Cristofaro, E. (2019) · Proceedings on Privacy Enhancing Technologies · doi:10.2478/popets-2019-0008 · Cited by 433
Train faster, generalize better: Stability of stochastic gradient descentPreprint
Hardt, M., Recht, B., & Singer, Y. (2015) · arXiv (Cornell University) · doi:10.48550/arxiv.1509.01240 · Cited by 431
Demystifying Membership Inference Attacks in Machine Learning as a Service
Truex, S., Liu, L., Gürsoy, M. E., Yu, L., & Wei, W. (2019) · IEEE Transactions on Services Computing · doi:10.1109/tsc.2019.2897554 · Cited by 421
Knock Knock, Who's There? Membership Inference on Aggregate Location Data
Pyrgelis, A., Troncoso, C., & Cristofaro, E. D. (2018) · Network and Distributed System Security Symposium · doi:10.14722/ndss.2018.23183 · Cited by 232
Making AI Forget You: Data Deletion in Machine Learning
Ginart, A., Guan, M. Y., Valiant, G., & Zou, J. (2019) · arXiv · doi:10.48550/arxiv.1907.05012 · Cited by 175
Zero-shot Knowledge Transfer via Adversarial Belief MatchingPreprint
Micaelli, P., & Storkey, A. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1905.09768 · Cited by 134
Critical Learning Periods in Deep Neural NetworksPreprint
Achille, A., Rovere, M., & Soatto, S. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1711.08856 · Cited by 112
Self-citationWhere is the Information in a Deep Neural Network?Preprint
Achille, A., Paolini, G., & Soatto, S. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1905.12213 · Cited by 84
Self-citationA PAC-Bayesian Tutorial with A Dropout BoundPreprint
McAllester, D. (2013) · arXiv (Cornell University) · doi:10.48550/arxiv.1307.2118 · Cited by 66
Golatkar, A., Achille, A., & Soatto, S. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1905.13277 · Cited by 54
Self-citationMachine UnlearningPreprint
Bourtoule, L., Chandrasekaran, V., Choquette-Choo, C. A., Jia, H., Travers, A., Zhang, B., Lie, D., & Papernot, N. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1912.03817 · Cited by 9
Flagged first, then by citations
2 of the 32 references couldn’t be matched to a known paper.
Research works printed but not linked: 2
- hardt, 2015, generalize better: Stability of stochastic gradient descent, ArXiv Preprint
- shintre, 2019, Verifying that the influence of a user data point has been removed from a machine learning classifier, US Patent App
Cited by
Sorted by citations
A Comprehensive Survey of Continual Learning: Theory, Method and Application
Wang, L., Zhang, X., Su, H., & Zhu, J. (2024) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2024.3367329 · Cited by 1,113
Bourtoule, L., Chandrasekaran, V., Choquette-Choo, C. A., Jia, H., Travers, A., Zhang, B., Lie, D., & Papernot, N. (2021) · 2021 IEEE Symposium on Security and Privacy (SP) · doi:10.1109/sp40001.2021.00019 · Cited by 810
Scope of machine learning in materials research—A review
Mobarak, M. H., Mimona, M. A., Islam, M. A., Hossain, N., Zohura, F. T., Imtiaz, I., & Rimon, M. I. H. (2023) · Applied Surface Science Advances · doi:10.1016/j.apsadv.2023.100523 · Cited by 249
Erasing Concepts from Diffusion Models
Gandikota, R., Materzyńska, J., Fiotto-Kaufman, J., & Bau, D. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.00230 · Cited by 355
FedEraser: Enabling Efficient Client-Level Data Removal from Federated Learning Models
Liu, G., Ma, X., Yang, Y., Wang, C., & Liu, J. (2021) · 2021 IEEE/ACM 29th International Symposium on Quality of Service (IWQOS) · doi:10.1109/iwqos52092.2021.9521274 · Cited by 248
Metadata
- Authors
- Golatkar, A., Achille, A., & Soatto, S.
- Affiliations
- University of California, Los Angeles (US)
- Venue
- 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
- Publisher
- IEEE
- Published
- 2020-06
- Pages
- 9301-9309
- ISSN
- 1063-6919 (ISSN-L)
- OpenAlex
- W2985940692
- Semantic Scholar
- c956d133b78e0d9b20885593809f0b636ce34093
- Cited by
- 394
- References
- 32
- Citations by year
- 2026: 96 · 2025: 115 · 2024: 75 · 2023: 56 · 2022: 24 · 2021: 21 · 2020: 6
- FWCI
- 10.22
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Licence
- Locations
- IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- Office of Naval Research (ONR N00014-17-1-2072, N00014-19-1-2229, ONR-N00014, N00014-17-1, N00014-19, N00014-19-1-2066, N00014-17-1-2072, N00014, N00014-17)
- Army Research Office (W911NF-17, W911NF-17-1-0304, W911NF, ARO W911NF-17-1-0304)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Privacy-Preserving Technologies in Data
- Physical Sciences › Computer Science › Artificial Intelligence › Adversarial Robustness in Machine Learning
- Physical Sciences › Computer Science › Computer Networks and Communications › Cryptography and Data Security
- Keywords
- selective forgetting, deep neural network, differential privacy, information erasure, information leakage, privacy-preserving machine learning
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
