The DOI matches its Crossref record; 8 of the 8 references it cites were found in a registry or an index.
Stochastic Gradient Descent Performs Variational Inference, Converges to Limit Cycles for Deep Networks
Open accessAbstract
Stochastic gradient descent (SGD) is widely believed to perform implicit regularization when used to train deep neural networks, but the precise manner in which this occurs has thus far been elusive. We prove that SGD minimizes an average potential over the posterior distribution of weights along with an entropic regularization term. This potential is however not the original loss function in general. So SGD does perform variational inference, but for a different loss than the one used to compute the gradients. Even more surprisingly, SGD does not even converge in the classical sense: we show that the most likely trajectories of SGD for deep networks do not behave like Brownian motion around critical points. Instead, they resemble closed loops with deterministic components. We prove that such “out-of-equilibrium” behavior is a consequence of highly nonisotropic gradient noise in SGD; the covariance matrix of mini-batch gradients for deep networks has a rank as small as 1% of its dimension. We provide extensive empirical validation of these claims. This article summarizes the findings in [1]. See the longer version for background, detailed results and proofs.
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.
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
Gradient-based learning applied to document recognition
Lecun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998) · Proceedings of the IEEE · doi:10.1109/5.726791 · Cited by 58,805
Neural Architecture Search with Reinforcement Learning
Zoph, B., & Le, Q. V. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1611.01578 · Cited by 3,974
An Introduction to Variational Methods for Graphical Models
Jordan, M. I., Ghahramani, Z., Jaakkola, T. S., & Saul, L. K. (1999) · Machine Learning · doi:10.1023/a:1007665907178 · Cited by 3,791
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
The information bottleneck method
Tishby, N., Pereira, F. C., & Bialek, W. (2000) · arXiv · doi:10.48550/arxiv.physics/0004057 · Cited by 2,626 · cited 2×
“The functional (11) is equivalent to the information bottleneck principle in representation learning [11].”
BackgroundAccurate, Large Minibatch SGD: Training ImageNet in 1 Hour
Goyal, P., Dollár, P., Girshick, R., Noordhuis, P., Wesolowski, L., Kyrola, A., Tulloch, A., Jia, Y., & He, K. (2017) · arXiv · doi:10.48550/arxiv.1706.02677 · Cited by 2,294
Keeping the neural networks simple by minimizing the description length of the weights
Hinton, G. E., & van Camp, D. (1993) · Proceedings of the sixth annual conference on Computational learning theory - COLT '93 · doi:10.1145/168304.168306 · Cited by 994
Beyond Equilibrium Thermodynamics
Öttinger, H. C. (2005) · doi:10.1002/0471727903 · Cited by 896
Cited by
A Simple Baseline for Bayesian Uncertainty in Deep Learning
Maddox, W., Garipov, T., Izmailov, P., Vetrov, D., & Wilson, A. G. (2019) · arXiv · doi:10.48550/arxiv.1902.02476 · Cited by 429
The Embedding Hypothesis: From Fourier Circuits to No-Q Attention
Rigoni, N. (2026) · Zenodo (CERN European Organization for Nuclear Research) · doi:10.5281/zenodo.19273886 · Cited by 292
Inductive biases for deep learning of higher-level cognition
Goyal, A., & Bengio, Y. (2022) · Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences · doi:10.1098/rspa.2021.0068 · Cited by 300
Implicit Gradient Regularization
Barrett, D. G. T., & Dherin, B. (2020) · arXiv · doi:10.48550/arxiv.2009.11162 · Cited by 95
Bayesian mechanics for stationary processes
Da Costa, L., Friston, K., Heins, C., & Pavliotis, G. A. (2021) · Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences · doi:10.1098/rspa.2021.0518 · Cited by 63
Metadata
- Authors
- Chaudhari, P., & Soatto, S.
- Venue
- 2018 Information Theory and Applications Workshop (ITA)
- Publisher
- IEEE
- Published
- 2018-02
- Pages
- 1-10
- OpenAlex
- W2765393161
- Cited by
- 213
- References
- 8
- FWCI
- 15.08
- 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
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Stochastic Gradient Descent Performs Variational Inference, Converges to Limit Cycles for Deep Networks. https://app.cochat.ai/papers/10.1109/ita.2018.8503224
- The paper
- Chaudhari, P., & Soatto, S. (2018). Stochastic Gradient Descent Performs Variational Inference, Converges to Limit Cycles for Deep Networks. 2018 Information Theory and Applications Workshop (ITA). https://doi.org/10.1109/ita.2018.8503224
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.