Papers › Understanding Approximate Fisher Information for Fast Convergence of Natural Gradient...

Understanding Approximate Fisher Information for Fast Convergence of Natural Gradient Descent in Wide Neural Networks

2 Oct 2020NeurIPS 2020 12arXiv:2010.00879archive 2025-07-28

Ryo Karakida, Kazuki Osawa

Natural Gradient Descent (NGD) helps to accelerate the convergence of gradient descent dynamics, but it requires approximations in large-scale deep neural networks because of its high computational cost. Empirical studies have confirmed that some NGD methods with approximate Fisher information converge sufficiently fast in practice. Nevertheless, it remains unclear from the theoretical perspective why and under what conditions such heuristic approximations work well. In this work, we reveal that, under specific conditions, NGD with approximate Fisher information achieves the same fast convergence to global minima as exact NGD. We consider deep neural networks in the infinite-width limit, and analyze the asymptotic training dynamics of NGD in function space via the neural tangent kernel. In the function space, the training dynamics with the approximate Fisher information are identical to those with the exact Fisher information, and they converge quickly. The fast convergence holds in layer-wise approximations; for instance, in block diagonal approximation where each block corresponds to a layer as well as in block tri-diagonal and K-FAC approximations. We also find that a unit-wise approximation achieves the same fast convergence under some assumptions. All of these different approximations have an isotropic gradient in the function space, and this plays a fundamental role in achieving the same convergence properties in training. Thus, the current study gives a novel and unified theoretical foundation with which to understand NGD methods in deep learning.

PaperPDFConference PDFCodeCode Syntology ran

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, on Syntology's MCP service (how to connect):

Code

Syntology 11 code samples harvested from 1 repository linked to this paper: 8 ran (of which 0 constructed an object rather than computing a result; 8 with no instrument failure: 0 honoured, 0 violated, 8 with no contract checked; 0 where Syntology's instrument failed) · 3 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code; here there are none.

By repository: official repository: 11 samples from 1 repository, 8 ran (8 with no instrument failure; 0 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

Levels written on 2026-09-25 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

kazukiosawa/ngd_in_wide_nn officialmentioned in papermentioned on GitHubjaxMITSyntology: 11 samples, 8 ran (0 instrument-failed) · 3 unverified report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

11 samples harvested: 8 ran (of which 0 constructed an object rather than computing a result; 8 with no instrument failure: 0 honoured, 0 violated, 8 with no contract checked; 0 where Syntology's instrument failed) · 3 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page has not changed since , the first build that kept a record date for it; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

8ran
3unverified

Licence: 0 of the 11 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from kazukiosawa/ngd_in_wide_nn. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of six reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity.

kazukiosawa/ngd_in_wide_nn official repository 11 samples, 8 ran (0 instrument-failed) · 3 unverified
activation in 30 papers numpy-based/empirical.py ran MIT (permissive) · 91f17369226529c0 · report fixture: signature inputs:Tensor/2/float, name:str/0/str · returned tuple · level recorded 2026-09-25
analytic_kernel in 6 papers numpy-based/analytical.py ran MIT (permissive) · 06a25a41e742052f · report fixture: signature x1:ndarray/2/float, x2:ndarray/2/float, w_var:float/0/float, ng_type:str/0/str · returned list · level recorded 2026-09-25
batch_analytic_kernel numpy-based/analytical.py ran MIT (permissive) · f43af3f41a794003 · report fixture: signature x1:ndarray/2/float, x2:ndarray/2/float, w_var:float/0/float, ng_type:str/0/str, batch_size:int/0/int · returned list · level recorded 2026-09-25
forster_transform numpy-based/utils.py ran MIT (permissive) · 2b21995a01e13c49 · report fixture: signature X:ndarray/2/float, n_iters:int/0/int, tol:float/0/float, eps:float/0/float · returned ndarray · level recorded 2026-09-25
forward_and_backward numpy-based/empirical.py ran MIT (permissive) · c07dde587e3e19a7 · report fixture: signature x:ndarray/2/float, W1:ndarray/2/float, W2:ndarray/2/float, W3:ndarray/2/float, backward:bool/0/bool, kfac:bool/0/bool · returned tuple · level recorded 2026-09-25
gradient_descent_mse numpy-based/predict.py ran MIT (permissive) · 12b893d866119885 · report fixture: signature fx_train:ndarray/1/float, y_train:ndarray/1/float, g_dd:ndarray/2/float, g_td:ndarray/2/float · returned function · level recorded 2026-09-25
init_ntk_mlp numpy-based/empirical.py ran MIT (permissive) · efc518313b8e690c · report fixture: signature M3:int/0/int, M2:int/0/int, M1:int/0/int, M0:int/0/int, W_std:float/0/float, b_std:float/0/float, random_seed:int/0/int · returned tuple · level recorded 2026-09-25
load_dataset in 276 papers jax-based/utils.py ran MIT (permissive) · 272b1343d9ce6f9b · report fixture: signature dataset_path:str/0/str, n_train:int/0/int, n_test:int/0/int, n_classes:int/0/int · returned tuple · level recorded 2026-09-25
exact_natural_gradient_descent_mse numpy-based/predict.py unverified MIT (permissive) · df49c5fd9bfb5a9c · report
flatten_features jax-based/empirical.py unverified MIT (permissive) · 0e413bb1862ac2de · report
natural_gradient_descent_mse numpy-based/predict.py unverified MIT (permissive) · d11096b872285991 · report

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections