Papers › Diverse Gaussian Noise Consistency Regularization for Robustness and Uncertainty Calibration

Diverse Gaussian Noise Consistency Regularization for Robustness and Uncertainty Calibration

2 Apr 2021arXiv:2104.01231archive 2025-07-28

Theodoros Tsiligkaridis, Athanasios Tsiligkaridis

Deep neural networks achieve high prediction accuracy when the train and test distributions coincide. In practice though, various types of corruptions occur which deviate from this setup and cause severe performance degradations. Few methods have been proposed to address generalization in the presence of unforeseen domain shifts. In particular, digital noise corruptions arise commonly in practice during the image acquisition stage and present a significant challenge for current methods. In this paper, we propose a diverse Gaussian noise consistency regularization method for improving robustness of image classifiers under a variety of corruptions while still maintaining high clean accuracy. We derive bounds to motivate and understand the behavior of our Gaussian noise consistency regularization using a local loss landscape analysis. Our approach improves robustness against unforeseen noise corruptions by 4.2-18.4% over adversarial training and other strong diverse data augmentation baselines across several benchmarks. Furthermore, it improves robustness and uncertainty calibration by 3.7% and 5.5%, respectively, against all common corruptions (weather, digital, blur, noise) when combined with state-of-the-art diverse data augmentations.

PaperPDFCodeCode Syntology ranMethod as code

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 52 code samples harvested from 1 repository linked to this paper: 45 ran (of which 8 constructed an object rather than computing a result; 20 with no instrument failure: 3 honoured, 1 violated, 16 with no contract checked; 25 where Syntology's instrument failed) · 7 unverified. Where Syntology's instrument failed: 23 where our contract draft was wrong; 2 where the contract check could not run; failures of Syntology's instrument, not of the code.

By repository: official repository: 52 samples from 1 repository, 45 ran (20 with no instrument failure; 25 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 between 2026-09-22 and 2026-09-25 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed. 4 of the 45 samples that ran have no recorded date.

theot1/dign officialmentioned in papermentioned on GitHubpytorchMITSyntology: 52 samples, 45 ran (25 instrument-failed) · 7 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

52 samples harvested: 45 ran (of which 8 constructed an object rather than computing a result; 20 with no instrument failure: 3 honoured, 1 violated, 16 with no contract checked; 25 where Syntology's instrument failed) · 7 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.

3ran · honoured contract
1ran · violated contract
23ran · our draft was wrong
2ran; the contract check could not run
16ran
7unverified

Licence: 1 of the 52 samples is pointer only, meaning Syntology does not serve that copy's text. For that sample, this paper's copy is under a permissive licence and cleared, but the stored copy of the same code that Syntology serves is not; the licence label's tooltip gives Syntology's recorded reason. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from theot1/dign. “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”. For 3 samples that ran, the graph records a signature this page does not show: for a class the fixture built without calling, the harness called the object with inputs it made up itself, so the recorded signature is the harness's probe, not the fixture's arguments. “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. A census tag appears when Syntology's census ran implementations with that name on one shared input per bucket: it gives how many were compared, from how many papers, and how many distinct outputs they produced, and names this sample's bucket when this sample was one of them.

theot1/dign official repository 52 samples, 45 ran (25 instrument-failed) · 7 unverified
AugMaxModule DiGN.py ran MIT (permissive) · 1e6ccf5662228a75 · report fixture: signature xs:tuple/1/Tensor, m:Tensor/1/float, q:Tensor/2/float · returned Tensor · level recorded 2026-09-24
AugMix DiGN.py ran; the contract check could not run (RAISES) MIT (permissive) · 9e4262567e4d68af · report fixture: returned Tensor · level recorded 2026-09-22
BasicBlock in 126 papers
census: 6 implementations from 6 papers, 5 distinct outputs on one shared input
DiGN.py ran · metamorphic tier: deterministic fingerprinted MIT (permissive) · 9fcfe39f01a6db07 · report fixture: returned BasicBlock · level recorded 2026-09-22
Bottle2neck DiGN.py ran MIT (permissive) · 386e139e40ce0e23 · report fixture: returned Bottle2neck · level recorded 2026-09-22
Bottleneck in 98 papers DiGN.py ran · metamorphic tier: deterministic fingerprinted MIT (permissive) · c5503c74c1ea7353 · report fixture: returned Bottleneck · level recorded 2026-09-22
DenseNet in 5 papers DiGN.py ran · metamorphic tier: deterministic MIT (permissive) · 757ed71343ce4c3b · report fixture: returned DenseNet · level recorded 2026-09-22
DenseNet121 in 32 papers DiGN.py ran · our draft was wrong MIT (permissive) · e18432664d73b55f · report fixture: signature num_classes:int/0/int · returned DenseNet · level recorded 2026-09-22
DenseNet121 in 32 papers densenet.py ran MIT (permissive) · 9ccdb4ad5a50fc5c · report fixture: signature num_classes:int/0/int, growth_rate:int/0/int, reduction:float/0/float · returned DenseNet · level recorded 2026-09-25
FakeReLU DiGN.py ran MIT (permissive) · 06935f102e16e291 · report fixture: returned FakeReLU · level recorded 2026-09-22
Res2Net DiGN.py ran MIT (permissive) · 2f365ca72da7a611 · report fixture: returned Res2Net · level recorded 2026-09-22
ResNet in 139 papers DiGN.py ran · metamorphic tier: deterministic MIT (permissive) · b3e481f3ef6524ff · report fixture: returned ResNet · level recorded 2026-09-22
ResNet18 in 174 papers DiGN.py ran · our draft was wrong MIT (permissive) · 6dfc13a796fd363a · report fixture: signature num_classes:int/0/int · returned ResNet · level recorded 2026-09-22
ResNet18 in 174 papers resnet.py ran MIT (permissive) · 1fb78623e6d1393a · report fixture: signature num_classes:int/0/int, feat_scale:int/0/int, wm:int/0/int, d_input:int/0/int · returned ResNet · level recorded 2026-09-25
ResNet18Wide DiGN.py ran · our draft was wrong MIT (permissive) · 09e581aa9c0b8963 · report fixture: signature num_classes:int/0/int · returned ResNet · level recorded 2026-09-22
ResNet18Wide resnet.py ran MIT (permissive) · 46fb144394e21588 · report fixture: signature num_classes:int/0/int · returned ResNet · level recorded 2026-09-25
ResNet18Wide_64 DiGN.py ran · our draft was wrong MIT (permissive) · 369c908be6ae465b · report fixture: signature num_classes:int/0/int · returned ResNet · level recorded 2026-09-22
ResNet18_64 DiGN.py ran · our draft was wrong MIT (permissive) · 970d810339028979 · report fixture: signature num_classes:int/0/int · returned ResNet · level recorded 2026-09-22
ResNet18_64 resnet.py ran MIT (permissive) · 7bc30cc42373d340 · report fixture: signature num_classes:int/0/int · returned ResNet · level recorded 2026-09-25
ResNet50 in 83 papers DiGN.py ran · our draft was wrong MIT (permissive) · a19e7b98eeb4b6cd · report fixture: signature num_classes:int/0/int · returned ResNet · level recorded 2026-09-22
Transition in 7 papers DiGN.py ran · metamorphic tier: deterministic fingerprinted MIT (permissive) · da9155d90cea93d6 · report fixture: returned Transition · level recorded 2026-09-22
apply_op in 5 papers DiGN.py ran · our draft was wrong MIT (permissive) · cc5cf9b07ce0d984 · report fixture: signature image:ndarray/2/float, op:callable/0/?, severity:float/0/float, d_input:dict/0/? · returned ndarray · level recorded 2026-09-22
augment in 36 papers DiGN.py ran · our draft was wrong fingerprinted MIT (permissive) · 269d965e70b2c96f · report fixture: signature image:ndarray/3/float, severity:int/0/int, width:int/0/int, depth:int/0/int · returned list · level recorded 2026-09-22
augment_and_mix DiGN.py ran; the contract check could not run (RAISES) · metamorphic tier: invariant fingerprinted MIT (permissive) · ec1ce1fd53588761 · report fixture: signature image:ndarray/3/float, severity:int/0/int, width:int/0/int, depth:int/0/int, alpha:float/0/float · returned ndarray · level recorded 2026-09-22
autocontrast in 9 papers DiGN.py ran · our draft was wrong MIT (permissive) · f9666ba13dd406de · report fixture: signature pil_img:Image/0/?, _:NoneType/0/?, IMAGE_SIZE:int/0/int · returned Image · level recorded 2026-09-22
calibr_metrics test-cal-c.py ran · our draft was wrong fingerprinted MIT (permissive) · 09d8f6afc431411e · report fixture: signature p_pred:ndarray/2/float, y_true:ndarray/1/int · returned tuple · level recorded, date not recorded
conv1x1 in 691 papers imagenet_models/leaky_resnet.py ran · our draft was wrong MIT (permissive) · d9def42110729a85 · report fixture: signature in_planes:int/0/int, out_planes:int/0/int, stride:int/0/int · returned Conv2d · level recorded, date not recorded
conv3x3 in 1,376 papers imagenet_models/resnet.py ran · our draft was wrong MIT (permissive) · 160bb14bd76201b4 · report fixture: signature in_planes:int/0/int, out_planes:int/0/int, stride:int/0/int, groups:int/0/int, dilation:int/0/int · returned Conv2d · level recorded, date not recorded
conv3x3 in 1,376 papers imagenet_models/leaky_resnet.py ran · our draft was wrong MIT (permissive) here; the stored copy is pointer only · fac5364e2f53c6db · report fixture: signature in_planes:int/0/int, out_planes:int/0/int, stride:int/0/int · returned Conv2d · level recorded, date not recorded
equalize in 10 papers DiGN.py ran · our draft was wrong MIT (permissive) · d1d727ae7d403f9d · report fixture: signature pil_img:Image/0/?, _:NoneType/0/?, IMAGE_SIZE:int/0/int · returned Image · level recorded 2026-09-22
float_parameter in 36 papers DiGN.py ran · honoured contract MIT (permissive) · 340cb3797f4ed00a · report fixture: signature level:int/0/int, maxval:float/0/float · returned float · level recorded 2026-09-22
genAugs DiGN.py ran · our draft was wrong MIT (permissive) · 6628458b28e57fbb · report fixture: returned list · level recorded 2026-09-22
get_loader_from_path DataLoad.py ran MIT (permissive) · 66d02694965ae38a · report fixture: signature path:str/0/str, batch_size:int/0/int, num_workers:int/0/int, d_input:int/0/int · returned DataLoader · level recorded 2026-09-25
get_loaders_imagenette DataLoad.py ran MIT (permissive) · 27fddf2dbd26fc4b · report fixture: signature data_path:str/0/str, batch_size_train:int/0/int, batch_size_val:int/0/int, num_workers:int/0/int · returned tuple · level recorded 2026-09-25
get_loaders_tinyimagenet DataLoad.py ran MIT (permissive) · 32bd96a059966be0 · report fixture: signature data_path:str/0/str, batch_size_train:int/0/int, batch_size_val:int/0/int, num_workers:int/0/int · returned tuple · level recorded 2026-09-25
int_parameter in 35 papers DiGN.py ran · honoured contract fingerprinted MIT (permissive) · 8b20dcbdf9e59b56 · report fixture: signature level:int/0/int, maxval:int/0/int · returned int · level recorded 2026-09-22
inv_normalize DiGN.py ran · our draft was wrong MIT (permissive) · 63aa40a3b7aafb93 · report fixture: signature tensor:list/1/Tensor, mean:list/1/float, std:list/1/float · returned list · level recorded 2026-09-22
normalize in 638 papers
census: 70 implementations from 82 papers, 51 distinct outputs across 11 buckets, one shared input per bucket this sample: bucket 7 of 11, arg 1: rank 2, kind float, dtype float32 · arg 2: rank 1, kind float, dtype float32 · arg 3: rank 1, kind float, dtype float32
DiGN.py ran · honoured contract fingerprinted MIT (permissive) · 12e0fe0c3812d359 · report fixture: signature tensor:Tensor/2/float, mean:Tensor/1/float, std:Tensor/1/float · returned Tensor · level recorded 2026-09-22
posterize in 7 papers DiGN.py ran · our draft was wrong MIT (permissive) · 7762e73a88022552 · report fixture: signature pil_img:Image/0/?, level:int/0/int, IMAGE_SIZE:int/0/int · returned Image · level recorded 2026-09-22
rotate in 56 papers DiGN.py ran · our draft was wrong MIT (permissive) · 427c26c8b0d1e858 · report fixture: signature pil_img:Image/0/?, level:int/0/int, IMAGE_SIZE:int/0/int · returned Image · level recorded 2026-09-22
sample_level in 16 papers augmentations.py ran · violated contract fingerprinted MIT (permissive) · 5cc8d4764ac07a35 · report fixture: signature n:int/0/int · returned float · level recorded 2026-09-22
shear_x in 43 papers DiGN.py ran · our draft was wrong MIT (permissive) · 8baad92985baea5d · report fixture: signature pil_img:Image/0/?, level:int/0/int, IMAGE_SIZE:int/0/int · returned Image · level recorded 2026-09-22
shear_y in 37 papers DiGN.py ran · our draft was wrong MIT (permissive) · 0af81fabe08c27e1 · report fixture: signature pil_img:Image/0/?, level:int/0/int, IMAGE_SIZE:int/0/int · returned Image · level recorded 2026-09-22
solarize in 15 papers DiGN.py ran · our draft was wrong MIT (permissive) · 58bdedec5152aaaa · report fixture: signature pil_img:Image/0/?, level:int/0/int, IMAGE_SIZE:int/0/int · returned Image · level recorded 2026-09-22
translate_x in 7 papers DiGN.py ran · our draft was wrong MIT (permissive) · bb8ddf80c6f21c90 · report fixture: signature pil_img:Image/0/?, level:int/0/int, IMAGE_SIZE:int/0/int · returned Image · level recorded 2026-09-22
translate_y in 7 papers DiGN.py ran · our draft was wrong MIT (permissive) · 6785662f742188b4 · report fixture: signature pil_img:Image/0/?, level:int/0/int, IMAGE_SIZE:int/0/int · returned Image · level recorded 2026-09-22
DiGN DiGN.py unverified MIT (permissive) · 7cdbd71b633d65b8 · report
SequentialWithArgs DiGN.py unverified MIT (permissive) · 528114b7552b6911 · report
densenet121 in 43 papers imagenet_models/densenet.py unverified MIT (permissive) · 85c4757ab3246b61 · report
densenet161 in 20 papers imagenet_models/densenet.py unverified MIT (permissive) · 6b70762b5419b176 · report
densenet169 in 42 papers imagenet_models/densenet.py unverified MIT (permissive) · 019686268f24b3b5 · report
leaky_resnet18 imagenet_models/leaky_resnet.py unverified MIT (permissive) · d5b851531e5607ed · report
resnet18 in 432 papers imagenet_models/resnet.py unverified MIT (permissive) · ba4c156b3db83275 · report

Method as code Syntology

What Syntology's graph records as this paper's claimed method, lifted into code: each assembly names its bricks (functions and classes), each with its own role and its own verification level. An assembly is not a check that the code implements the method correctly or reproduces the paper's results. Levels are each brick's own verification level in Syntology's graph (level 0: no recorded run), listed brick by brick and never combined into one level; “ran” means the brick executed on a synthesized input, not that its output is correct, and the instrument figure counts failures of Syntology's instrument, not of the code. Where a model read the code against the paper's claim, each assembly's label verdict is one model's judgement, kept in this page's JSON twin; it is not calibrated, so no card prints it.

1 assembly on this paper: 1 harvested, 0 generated.

harvested Assembly for Diverse Gaussian Noise Consistency Regularization (graph method: Diverse Gaussian Noise Consistency Regularization). 37 bricks; individually on a fixture: 35 ran (of which 8 constructed an object rather than computing a result; 14 with no instrument failure: 4 honoured, 1 violated, 9 with no contract checked; 21 where Syntology's instrument failed) · 2 unverified. The assembly has not been run end-to-end.

  1. DiGN entry · level 0 · unverified · 7cdbd71b633d65b8
  2. AugMaxModule helper · level 1 · ran · 1e6ccf5662228a75
  3. AugMix helper · level 1 · ran; the contract check could not run (RAISES) · 9e4262567e4d68af
  4. DenseNet121 helper · level 1 · ran · our draft was wrong · e18432664d73b55f
  5. Res2Net helper · level 1 · ran · 2f365ca72da7a611
  6. ResNet18 helper · level 1 · ran · our draft was wrong · 6dfc13a796fd363a
  7. ResNet18Wide helper · level 1 · ran · our draft was wrong · 09e581aa9c0b8963
  8. ResNet18Wide_64 helper · level 1 · ran · our draft was wrong · 369c908be6ae465b
  9. ResNet18_64 helper · level 1 · ran · our draft was wrong · 970d810339028979
  10. ResNet50 helper · level 1 · ran · our draft was wrong · a19e7b98eeb4b6cd
  11. genAugs helper · level 1 · ran · our draft was wrong · 6628458b28e57fbb
  12. inv_normalize helper · level 1 · ran · our draft was wrong · 63aa40a3b7aafb93
  13. normalize helper · level 1 · ran · honoured contract · 12e0fe0c3812d359
  14. norms helper · level 1 · ran · honoured contract · b7ae66630d9857bc
  15. augment_and_mix helper · level 1 · ran; the contract check could not run (RAISES) · ec1ce1fd53588761
  16. Bottleneck helper · level 1 · ran · c5503c74c1ea7353
  17. DenseNet helper · level 1 · ran · 757ed71343ce4c3b
  18. Bottle2neck helper · level 1 · ran · 386e139e40ce0e23
  19. BasicBlock helper · level 1 · ran · 9fcfe39f01a6db07
  20. ResNet helper · level 1 · ran · b3e481f3ef6524ff
  21. augment helper · level 1 · ran · our draft was wrong · 269d965e70b2c96f
  22. apply_op helper · level 1 · ran · our draft was wrong · cc5cf9b07ce0d984
  23. FakeReLU helper · level 1 · ran · 06935f102e16e291
  24. Transition helper · level 1 · ran · da9155d90cea93d6
  25. SequentialWithArgs helper · level 0 · unverified · 528114b7552b6911
  26. autocontrast helper · level 1 · ran · our draft was wrong · f9666ba13dd406de
  27. equalize helper · level 1 · ran · our draft was wrong · d1d727ae7d403f9d
  28. float_parameter helper · level 1 · ran · honoured contract · 340cb3797f4ed00a
  29. int_parameter helper · level 1 · ran · honoured contract · 8b20dcbdf9e59b56
  30. posterize helper · level 1 · ran · our draft was wrong · 7762e73a88022552
  31. rotate helper · level 1 · ran · our draft was wrong · 427c26c8b0d1e858
  32. sample_level helper · level 1 · ran · violated contract · 5cc8d4764ac07a35
  33. shear_x helper · level 1 · ran · our draft was wrong · 8baad92985baea5d
  34. shear_y helper · level 1 · ran · our draft was wrong · 0af81fabe08c27e1
  35. solarize helper · level 1 · ran · our draft was wrong · 58bdedec5152aaaa
  36. translate_x helper · level 1 · ran · our draft was wrong · bb8ddf80c6f21c90
  37. translate_y helper · level 1 · ran · our draft was wrong · 6785662f742188b4

Each brick ends with its code_sha256, Syntology's identity for that exact code; click one to copy the get_code call that returns it. For agents, on Syntology's MCP service: assembly(arxiv_id="2104.01231") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

Data AugmentationImage Classification

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