Proof-of-Learning: Definitions and Practice
The DOI matches its Crossref record; 20 of the 20 references it cites were found in a registry or an index.
Abstract
Training machine learning (ML) models typically involves expensive iterative optimization. Once the model’s final parameters are released, there is currently no mechanism for the entity which trained the model to prove that these parameters were indeed the result of this optimization procedure. Such a mechanism would support security of ML applications in several ways. For instance, it would simplify ownership resolution when multiple parties contest ownership of a specific model. It would also facilitate the distributed training across untrusted workers where Byzantine workers might otherwise mount a denial-ofservice by returning incorrect model updates.In this paper, we remediate this problem by introducing the concept of proof-of-learning in ML. Inspired by research on both proof-of-work and verified computations, we observe how a seminal training algorithm, stochastic gradient descent, accumulates secret information due to its stochasticity. This produces a natural construction for a proof-of-learning which demonstrates that a party has expended the compute require to obtain a set of model parameters correctly. In particular, our analyses and experiments show that an adversary seeking to illegitimately manufacture a proof-of-learning needs to perform at least as much work than is needed for gradient descent itself.We also instantiate a concrete proof-of-learning mechanism in both of the scenarios described above. In model ownership resolution, it protects the intellectual property of models released publicly. In distributed training, it preserves availability of the training procedure. Our empirical evaluation validates that our proof-of-learning mechanism is robust to variance induced by the hardware (e.g., ML accelerators) and software stacks.
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
Learning representations by back-propagating errors
Rumelhart, D. E., Hinton, G. E., & Williams, R. J. (1986) · Nature · doi:10.1038/323533a0 · Cited by 31,892
SciPy 1.0: fundamental algorithms for scientific computing in Python
Virtanen, P., Gommers, R., Oliphant, T. E., Haberland, M., Reddy, T., Cournapeau, D., Burovski, E., Peterson, P., Weckesser, W., Bright, J., et al. (2019) · Apollo - University of Cambridge Repository · doi:10.17863/cam.64134 · Cited by 10,460
A Stochastic Approximation Method
Robbins, H., & Monro, S. (1951) · The Annals of Mathematical Statistics · doi:10.1214/aoms/1177729586 · Cited by 9,643
Intriguing properties of neural networks
Szegedy, C., Zaremba, W., Sutskever, I., Bruna, J., Erhan, D., Goodfellow, I., & Fergus, R. (2013) · arXiv · doi:10.48550/arxiv.1312.6199 · Cited by 8,143
The Kolmogorov-Smirnov Test for Goodness of Fit
Massey, F. J. (1951) · Journal of the American Statistical Association · doi:10.1080/01621459.1951.10500769 · Cited by 7,210
How transferable are features in deep neural networks?
Yosinski, J., Clune, J., Bengio, Y., & Lipson, H. (2014) · arXiv · doi:10.48550/arxiv.1411.1792 · Cited by 6,061 · cited 1×
“To this end, the adversary can fine-tune [67] or fine-prune [82] the model fWT to achieve f which is not an exact copy of fWT but has comparable test-time performance.”
BackgroundIn-Datacenter Performance Analysis of a Tensor Processing Unit
Jouppi, N. P., Young, C., Patil, N., Patterson, D., Agrawal, G., Bajwa, R., Bates, S., Bhatia, S., Boden, N., Borchers, A., et al. (2017) · Proceedings of the 44th Annual International Symposium on Computer Architecture · doi:10.1145/3079856.3080246 · Cited by 4,537 · cited 2×
“GPUs [2], TPUs [3], and FPGAs [4] are used to ensure efficient training.”
MethodPractical 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
Self-citationSelf-Normalizing Neural Networks
Klambauer, G., Unterthiner, T., Mayr, A., & Hochreiter, S. (2017) · arXiv · doi:10.48550/arxiv.1706.02515 · Cited by 2,267
Cited by
Unrolling SGD: Understanding Factors Influencing Machine Unlearning
Thudi, A., Deza, G., Chandrasekaran, V., & Papernot, N. (2022) · 2022 IEEE 7th European Symposium on Security and Privacy (EuroS&P) · doi:10.1109/eurosp53844.2022.00027 · Cited by 135
“Another issue is that one can retrain on the same sequence of data from the same initialization and obtain different terminal weights [23] (which is caused by randomness and numerical instabilities in floating point operations).”
BackgroundSelf-citationThe alignment problem from a deep learning perspective
Ngo, R., Chan, L., & Mindermann, S. (2022) · arXiv · doi:10.48550/arxiv.2209.00626 · Cited by 68
“Relevant work includes the development of proof-of-learning mechanisms to verify properties of training runs [Jia et al., 2021], tamper-resistant chip-level mechanisms (such as those on Nvidia’s Lite Hash Rate GPUs), and evaluation suites for dangerous capabilities.”
BackgroundA Survey of Machine Unlearning
Nguyen, T. T., Huynh, T. T., Ren, Z., Nguyen, P. L., Liew, A. W.-C., Yin, H., & Nguyen, Q. V. H. (2022) · arXiv · doi:10.48550/arxiv.2209.02299 · Cited by 193
“( ′ | 𝑤 ) is assumed to be a As a result, a candidate set of model parameters 𝑃𝑟 ( 𝑤 | 𝐷 ) is constructed by simply maximizing the posterior probability 𝑃𝑟 ( 𝑤 | 𝐷 ) , i.e.:”
zkDL: Efficient Zero-Knowledge Proofs of Deep Learning Training
Sun, H., Bai, T., Li, J., & Zhang, H. (2025) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2024.3520863 · Cited by 33
TruVRF: Toward Triple-Granularity Verification on Machine Unlearning
Zhou, C., Gao, Y., Fu, A., Chen, K., Zhang, Z., Xue, M., Dai, Z., Ji, S., & Zhang, Y. (2025) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2025.3565991 · Cited by 7
Metadata
- Authors
- Jia, H., Yaghini, M., Choquette-Choo, C. A., Dullerud, N., Thudi, A., Chandrasekaran, V., & Papernot, N.
- Venue
- 2021 IEEE Symposium on Security and Privacy (SP)
- Publisher
- IEEE
- Published
- 2021-05
- Pages
- 1039-1056
- ISSN
- 1081-6011 (ISSN-L)
- OpenAlex
- W3154581591
- Cited by
- 75
- References
- 20
- FWCI
- 5.18
- Type
- conference paper
- Language
- en
- Open access
- closed
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
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Proof-of-Learning: Definitions and Practice. https://app.cochat.ai/papers/10.1109/sp40001.2021.00106
- The paper
- Jia, H., Yaghini, M., Choquette-Choo, C. A., Dullerud, N., Thudi, A., Chandrasekaran, V., & Papernot, N. (2021). Proof-of-Learning: Definitions and Practice. 2021 IEEE Symposium on Security and Privacy (SP). https://doi.org/10.1109/sp40001.2021.00106
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.