Papers › ShortcutProbe: Probing Prediction Shortcuts for Learning Robust Models

ShortcutProbe: Probing Prediction Shortcuts for Learning Robust Models

20 May 2025arXiv:2505.13910archive 2025-07-28

Guangtao Zheng, Wenqian Ye, Aidong Zhang

Deep learning models often achieve high performance by inadvertently learning spurious correlations between targets and non-essential features. For example, an image classifier may identify an object via its background that spuriously correlates with it. This prediction behavior, known as spurious bias, severely degrades model performance on data that lacks the learned spurious correlations. Existing methods on spurious bias mitigation typically require a variety of data groups with spurious correlation annotations called group labels. However, group labels require costly human annotations and often fail to capture subtle spurious biases such as relying on specific pixels for predictions. In this paper, we propose a novel post hoc spurious bias mitigation framework without requiring group labels. Our framework, termed ShortcutProbe, identifies prediction shortcuts that reflect potential non-robustness in predictions in a given model's latent space. The model is then retrained to be invariant to the identified prediction shortcuts for improved robustness. We theoretically analyze the effectiveness of the framework and empirically demonstrate that it is an efficient and practical tool for improving a model's robustness to spurious bias on diverse datasets.

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 27 code samples harvested from 1 repository linked to this paper: 21 ran (of which 12 constructed an object rather than computing a result; 15 with no instrument failure: 1 honoured, 0 violated, 14 with no contract checked; 6 where Syntology's instrument failed) · 6 unverified. Where Syntology's instrument failed: 6 where our contract draft was wrong; failures of Syntology's instrument, not of the code.

By repository: official repository: 27 samples from 1 repository, 21 ran (15 with no instrument failure; 6 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 2026-09-22 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

gtzheng/shortcutprobe officialmentioned in paperpytorchSyntology: 27 samples, 21 ran (6 instrument-failed) · 6 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

27 samples harvested: 21 ran (of which 12 constructed an object rather than computing a result; 15 with no instrument failure: 1 honoured, 0 violated, 14 with no contract checked; 6 where Syntology's instrument failed) · 6 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.

1ran · honoured contract
6ran · our draft was wrong
14ran
6unverified

Licence: 27 of the 27 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 gtzheng/shortcutprobe. “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 1 sample 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.

gtzheng/shortcutprobe official repository 27 samples, 21 ran (6 instrument-failed) · 6 unverified
ClassBalancedSampler algorithms/shortcut_probe.py ran no licence file found · pointer only · 6e663667ba7ed3f2 · report fixture: returned ClassBalancedSampler · level recorded 2026-09-22
EmbedDataset algorithms/shortcut_probe.py ran no licence file found · pointer only · a87713c5b6ae5b15 · report fixture: returned EmbedDataset · level recorded 2026-09-22
GroupBalancedSampler algorithms/shortcut_probe.py ran no licence file found · pointer only · c468389298284a04 · report fixture: returned GroupBalancedSampler · level recorded 2026-09-22
IdxDataset algorithms/shortcut_probe.py ran · metamorphic tier: deterministic fingerprinted no licence file found · pointer only · 4711b44eb7381e64 · report fixture: signature idx:int/0/int · returned tuple · level recorded 2026-09-22
JointDataset algorithms/shortcut_probe.py ran no licence file found · pointer only · fa1c55aa70701d1e · report fixture: returned JointDataset · level recorded 2026-09-22
JointRandomSampler algorithms/shortcut_probe.py ran no licence file found · pointer only · a5207b840e3096a2 · report fixture: returned JointRandomSampler · level recorded 2026-09-22
MisDataset algorithms/shortcut_probe.py ran no licence file found · pointer only · 52fb87f5341104be · report fixture: returned MisDataset · level recorded 2026-09-22
PredDataset algorithms/shortcut_probe.py ran no licence file found · pointer only · 2ee64f5afd17a5a3 · report fixture: returned PredDataset · level recorded 2026-09-22
RandomSampler algorithms/shortcut_probe.py ran no licence file found · pointer only · 36087249bd727695 · report fixture: returned RandomSampler · level recorded 2026-09-22
RunningVariance algorithms/shortcut_probe.py ran no licence file found · pointer only · 4a0b5cf2bf60d9e7 · report fixture: returned RunningVariance · level recorded 2026-09-22
ShortcutSampler algorithms/shortcut_probe.py ran no licence file found · pointer only · f6cdc99333d7ac0c · report fixture: returned ShortcutSampler · level recorded 2026-09-22
SpuriousVectors algorithms/shortcut_probe.py ran fingerprinted no licence file found · pointer only · b3e59759edda6308 · report fixture: returned SpuriousVectors · level recorded 2026-09-22
SubsetDataset algorithms/shortcut_probe.py ran no licence file found · pointer only · 3be6f85f9a2c50c3 · report fixture: returned SubsetDataset · level recorded 2026-09-22
bert_adamw_optimizer algorithms/shortcut_probe.py ran · our draft was wrong no licence file found · pointer only · abb051ad54c04456 · report fixture: signature model:callable/0/?, lr:float/0/float, weight_decay:float/0/float · returned AdamW · level recorded 2026-09-22
disentangle_pca algorithms/shortcut_probe.py ran · honoured contract no licence file found · pointer only · 227d71f4a9526c48 · report fixture: signature embedding_dir:str/0/str, split:str/0/str, keep_ndims:int/0/int · returned ndarray · level recorded 2026-09-22
extract_feature_info algorithms/shortcut_probe.py ran · our draft was wrong no licence file found · pointer only · ffd024417d2e0ee9 · report fixture: signature model:callable/0/?, loader:SimpleLoader/0/?, embedding_dir:str/0/str, split:str/0/str, threshold:int/0/int, keep_ndims:int/0/int, device:str/0/str · returned tuple · level recorded 2026-09-22
get_split_indexes algorithms/shortcut_probe.py ran no licence file found · pointer only · 24df1f39ca418afa · report fixture: signature dataset:SimpleDataset/0/?, ratio:float/0/float · returned tuple · level recorded 2026-09-22
init_optimizer in 11 papers algorithms/shortcut_probe.py ran · our draft was wrong no licence file found · pointer only · ad87b4581c72eb68 · report fixture: signature model:callable/0/?, opt_name:str/0/str, kwargs:dict/0/? · returned AdamW · level recorded 2026-09-22
init_scheduler algorithms/shortcut_probe.py ran · our draft was wrong no licence file found · pointer only · 97a6fbbb8298e7a0 · report fixture: signature optimizer:SGD/0/?, scheduler_name:str/0/str, kwargs:dict/0/? · returned CosineAnnealingLR · level recorded 2026-09-22
load_embeddings in 33 papers algorithms/shortcut_probe.py ran · our draft was wrong no licence file found · pointer only · 2a04614f7104524c · report fixture: signature path:str/0/str, split:str/0/str, keep_ndims:int/0/int · returned tuple · level recorded 2026-09-22
register_algorithm algorithms/shortcut_probe.py ran · our draft was wrong no licence file found · pointer only · 6f906e5668960d23 · report fixture: signature name:str/0/str · returned function · level recorded 2026-09-22
Algorithm in 26 papers algorithms/shortcut_probe.py unverified no licence file found · pointer only · 9c581ceeab1e7194 · report
Classifier in 72 papers algorithms/shortcut_probe.py unverified no licence file found · pointer only · aa528f9aa86d7413 · report
Scaler algorithms/shortcut_probe.py unverified no licence file found · pointer only · 6a3dd1d39f936e97 · report
get_backbone in 28 papers algorithms/shortcut_probe.py unverified no licence file found · pointer only · 3db8286f1e2584d6 · report
prepare_data in 87 papers algorithms/shortcut_probe.py unverified no licence file found · pointer only · 0d22a2219be7c189 · report
shortcutprobe algorithms/shortcut_probe.py unverified no licence file found · pointer only · d99a2ddd28ee9061 · 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 ShortcutProbe (graph method: ShortcutProbe). 27 bricks; individually on a fixture: 21 ran (of which 12 constructed an object rather than computing a result; 15 with no instrument failure: 1 honoured, 0 violated, 14 with no contract checked; 6 where Syntology's instrument failed) · 6 unverified. The assembly has not been run end-to-end.

  1. shortcutprobe entry · level 0 · unverified · d99a2ddd28ee9061
  2. Algorithm helper · level 0 · unverified · 9c581ceeab1e7194
  3. Classifier helper · level 0 · unverified · aa528f9aa86d7413
  4. RunningVariance helper · level 1 · ran · 4a0b5cf2bf60d9e7
  5. SpuriousVectors model · level 1 · ran · b3e59759edda6308
  6. init_optimizer helper · level 1 · ran · our draft was wrong · ad87b4581c72eb68
  7. init_scheduler helper · level 1 · ran · our draft was wrong · 97a6fbbb8298e7a0
  8. register_algorithm helper · level 1 · ran · our draft was wrong · 6f906e5668960d23
  9. ClassBalancedSampler helper · level 1 · ran · 6e663667ba7ed3f2
  10. EmbedDataset helper · level 1 · ran · a87713c5b6ae5b15
  11. GroupBalancedSampler helper · level 1 · ran · c468389298284a04
  12. IdxDataset helper · level 1 · ran · 4711b44eb7381e64
  13. JointDataset helper · level 1 · ran · fa1c55aa70701d1e
  14. JointRandomSampler helper · level 1 · ran · a5207b840e3096a2
  15. MisDataset helper · level 1 · ran · 52fb87f5341104be
  16. PredDataset helper · level 1 · ran · 2ee64f5afd17a5a3
  17. RandomSampler helper · level 1 · ran · 36087249bd727695
  18. ShortcutSampler helper · level 1 · ran · f6cdc99333d7ac0c
  19. SubsetDataset helper · level 1 · ran · 3be6f85f9a2c50c3
  20. extract_feature_info helper · level 1 · ran · our draft was wrong · ffd024417d2e0ee9
  21. prepare_data helper · level 0 · unverified · 0d22a2219be7c189
  22. get_backbone helper · level 0 · unverified · 3db8286f1e2584d6
  23. bert_adamw_optimizer helper · level 1 · ran · our draft was wrong · abb051ad54c04456
  24. Scaler helper · level 0 · unverified · 6a3dd1d39f936e97
  25. disentangle_pca helper · level 1 · ran · honoured contract · 227d71f4a9526c48
  26. load_embeddings helper · level 1 · ran · our draft was wrong · 2a04614f7104524c
  27. get_split_indexes helper · level 1 · ran · 24df1f39ca418afa

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="2505.13910") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

Prediction

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

HOC

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