The DOI matches its DataCite record; 1 of the 1 references it cites were found in a registry or an index.
On the Implicit Bias of Adam
Abstract
In previous literature, backward error analysis was used to find ordinary differential equations (ODEs) approximating the gradient descent trajectory. It was found that finite step sizes implicitly regularize solutions because terms appearing in the ODEs penalize the two-norm of the loss gradients. We prove that the existence of similar implicit regularization in RMSProp and Adam depends on their hyperparameters and the training stage, but with a different "norm" involved: the corresponding ODE terms either penalize the (perturbed) one-norm of the loss gradients or, conversely, impede its reduction (the latter case being typical). We also conduct numerical experiments and discuss how the proven facts can influence generalization.
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?
In the citing paper
“Their focus is not on memory but on continuous approximations; in particular, they do not have approximation bounds between two discrete iterations like we do.”
From How Memory in Optimization Algorithms Implicitly Modifies the Loss (2025) · cites this paper as [11]
In the citing paper
“In addition, their arguments are highly specialized to Adam, and they do not incorporate weight decay.”
From How Memory in Optimization Algorithms Implicitly Modifies the Loss (2025) · cites this paper as [11]
In the citing paper
“As predicted in Section 4 and confirming the observation from [11] for pure Adam without weight decay, the test accuracy drops as β 2 approaches one.”
From How Memory in Optimization Algorithms Implicitly Modifies the Loss (2025) · cites this paper as [11]
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
When Vision Transformers Outperform ResNets without Pre-training or Strong Data Augmentations
Chen, X., Hsieh, C.-J., & Gong, B. (2021) · arXiv · doi:10.48550/arxiv.2106.01548 · Cited by 128
Cited by
How Memory in Optimization Algorithms Implicitly Modifies the Loss
Cattaneo, M., & Shigida, B. (2025) · Advances in Neural Information Processing Systems 38 · doi:10.52202/085713-5215
Self-citation
Metadata
- Authors
- Cattaneo, M. D., Klusowski, J. M., & Shigida, B.
- Venue
- arXiv
- Publisher
- arXiv
- Published
- 2023
- arXiv
- 2309.00079
- References
- 1
- Type
- Preprint
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: On the Implicit Bias of Adam. https://app.cochat.ai/papers/10.48550/arxiv.2309.00079
- The paper
- Cattaneo, M. D., Klusowski, J. M., & Shigida, B. (2023). On the Implicit Bias of Adam. arXiv. https://doi.org/10.48550/arxiv.2309.00079
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.