Papers › Generalization in diffusion models arises from geometry-adaptive harmonic representations

Generalization in diffusion models arises from geometry-adaptive harmonic representations

4 Oct 2023arXiv:2310.02557archive 2025-07-28

Zahra Kadkhodaie, Florentin Guth, Eero P. Simoncelli, Stéphane Mallat

Deep neural networks (DNNs) trained for image denoising are able to generate high-quality samples with score-based reverse diffusion algorithms. These impressive capabilities seem to imply an escape from the curse of dimensionality, but recent reports of memorization of the training set raise the question of whether these networks are learning the "true" continuous density of the data. Here, we show that two DNNs trained on non-overlapping subsets of a dataset learn nearly the same score function, and thus the same density, when the number of training images is large enough. In this regime of strong generalization, diffusion-generated images are distinct from the training set, and are of high visual quality, suggesting that the inductive biases of the DNNs are well-aligned with the data density. We analyze the learned denoising functions and show that the inductive biases give rise to a shrinkage operation in a basis adapted to the underlying image. Examination of these bases reveals oscillating harmonic structures along contours and in homogeneous regions. We demonstrate that trained denoisers are inductively biased towards these geometry-adaptive harmonic bases since they arise not only when the network is trained on photographic images, but also when it is trained on image classes supported on low-dimensional manifolds for which the harmonic basis is suboptimal. Finally, we show that when trained on regular image classes for which the optimal basis is known to be geometry-adaptive and harmonic, the denoising performance of the networks is near-optimal.

PaperPDFCodeCode 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 16 code samples harvested from 1 repository linked to this paper: 11 ran (of which 0 constructed an object rather than computing a result; 11 with no instrument failure: 0 honoured, 0 violated, 11 with no contract checked; 0 where Syntology's instrument failed) · 5 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code; here there are none.

By repository: official repository: 16 samples from 1 repository, 11 ran (11 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 2026-09-24 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

labforcomputationalvision/memorization_generalization_in_diffusion_models officialmentioned in papermentioned on GitHubpytorchMITSyntology: 16 samples, 11 ran (0 instrument-failed) · 5 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

16 samples harvested: 11 ran (of which 0 constructed an object rather than computing a result; 11 with no instrument failure: 0 honoured, 0 violated, 11 with no contract checked; 0 where Syntology's instrument failed) · 5 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.

11ran
5unverified

Licence: 0 of the 16 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 labforcomputationalvision/memorization_generalization_in_diffusion_models. “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.

labforcomputationalvision/memorization_generalization_in_diffusion_models official repository 16 samples, 11 ran (0 instrument-failed) · 5 unverified
MSE in 28 papers code/quality_metrics_func.py ran fingerprinted MIT (permissive) · eb15c404ffac080f · report fixture: signature ref_im:Tensor/4/float, im:Tensor/4/float · returned Tensor · level recorded 2026-09-24
batch_psnr_numpy code/quality_metrics_func.py ran MIT (permissive) · 1aa782be31b7fc00 · report fixture: signature ref_ims:ndarray/3/float, images:ndarray/3/float, max_I:float/0/float · returned ndarray · level recorded 2026-09-24
calc_jacobian code/linear_approx.py ran MIT (permissive) · d0324c885182c509 · report fixture: signature inp:Tensor/4/float, model:callable/0/?, layer_num:NoneType/0/?, channel_num:NoneType/0/? · returned Tensor · level recorded 2026-09-24
calc_jacobian_rows code/linear_approx.py ran MIT (permissive) · 98dacb3ad56237ba · report fixture: signature inp:Tensor/4/float, model:callable/0/?, i:int/0/int, j:int/0/int · returned Tensor · level recorded 2026-09-24
load_mnist in 35 papers code/dataloader_func.py ran MIT (permissive) · 6ad8ccbe18be47c9 · report fixture: signature data_dir_path:str/0/str · returned tuple · level recorded 2026-09-24
load_set14 code/dataloader_func.py ran MIT (permissive) · 31ac25abb6e95520 · report fixture: returned ndarray · level recorded 2026-09-24
make_fft_filter_1d code/synthetic_data_generators.py ran fingerprinted MIT (permissive) · 6eb2528b9343169a · report fixture: signature N:int/0/int · returned Tensor · level recorded 2026-09-24
make_fft_filter_2d code/synthetic_data_generators.py ran fingerprinted MIT (permissive) · faa53db8ba9db059 · report fixture: signature N:int/0/int · returned Tensor · level recorded 2026-09-24
make_fft_filter_2d_seprable code/synthetic_data_generators.py ran fingerprinted MIT (permissive) · 56aaa6b239a0ae60 · report fixture: signature N:int/0/int · returned Tensor · level recorded 2026-09-24
sig_to_psnr code/quality_metrics_func.py ran fingerprinted MIT (permissive) · 7ff802a286a451e5 · report fixture: signature std:float/0/float, I_max:int/0/int · returned float64 · level recorded 2026-09-24
single_image_loader code/dataloader_func.py ran MIT (permissive) · bf7b787bf71b1055 · report fixture: signature data_set_dire_path:str/0/str, image_number:int/0/int · returned Tensor · level recorded 2026-09-24
calc_jacobian_row_MS code/linear_approx.py unverified MIT (permissive) · 53ef977c62a678b6 · report
init_BF_CNN_RF code/model_loader_func.py unverified MIT (permissive) · 561901e6303b8afc · report
init_UNet code/model_loader_func.py unverified MIT (permissive) · 0a2d1378af0b3c3c · report
initialize_network code/model_loader_func.py unverified MIT (permissive) · 706dea6429c294c3 · report
univ_inv_sol code/algorithm_inv_prob.py unverified MIT (permissive) · 4835e614b7b50339 · report

Tasks

DenoisingImage DenoisingMemorization

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

Diffusion

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