Papers › Learning Optimal Conformal Classifiers

Learning Optimal Conformal Classifiers

18 Oct 2021ICLR 2022 4arXiv:2110.09192archive 2025-07-28

David Stutz, Krishnamurthy, Dvijotham, Ali Taylan Cemgil, Arnaud Doucet

Modern deep learning based classifiers show very high accuracy on test data but this does not provide sufficient guarantees for safe deployment, especially in high-stake AI applications such as medical diagnosis. Usually, predictions are obtained without a reliable uncertainty estimate or a formal guarantee. Conformal prediction (CP) addresses these issues by using the classifier's predictions, e.g., its probability estimates, to predict confidence sets containing the true class with a user-specified probability. However, using CP as a separate processing step after training prevents the underlying model from adapting to the prediction of confidence sets. Thus, this paper explores strategies to differentiate through CP during training with the goal of training model with the conformal wrapper end-to-end. In our approach, conformal training (ConfTr), we specifically "simulate" conformalization on mini-batches during training. Compared to standard training, ConfTr reduces the average confidence set size (inefficiency) of state-of-the-art CP methods applied after training. Moreover, it allows to "shape" the confidence sets predicted at test time, which is difficult for standard CP. On experiments with several datasets, we show ConfTr can influence how inefficiency is distributed across classes, or guide the composition of confidence sets in terms of the included classes, while retaining the guarantees offered by CP.

PaperPDFConference PDFCodeCode 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.

Syntology holds a reference implementation of Conformal Training (ConfTr), generated from the paper text (not the authors' code), at level 3 on Syntology's ladder: V2 “two implementations GENERATED from the paper's signature and spec agree numerically (1e-6) on fixture inputs” “neither of the two ever read the authors' own code, so V2 measures how unambiguously the paper specifies the method, not fidelity to the authors' implementation (measured 2026-09-11, commit 75eaa77: cross_checked_by is a second LLM on 2,628 of 2,831 served samples)”; V3 “paper-derived property tests pass on both of those same two generations.” (quoted from Syntology's server card). No rung compares this code against the authors' implementation. For agents: get_reference_implementation(name='Conformal Training (ConfTr)') (how to connect); the free check have('2110.09192') needs no token.

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

Code

Syntology 24 code samples harvested from 2 repositories linked to this paper: 18 ran (of which 0 constructed an object rather than computing a result; 18 with no instrument failure: 1 honoured, 0 violated, 17 with no contract checked; 0 where Syntology's instrument failed) · 6 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code; here there are none.

By repository: official repository: 21 samples from 1 repository, 17 ran (17 with no instrument failure; 0 where Syntology's instrument failed); community (archive-listed): 3 samples from 1 repository, 1 ran (1 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 between 2026-09-22 and 2026-09-24 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

deepmind/conformal_training officialjaxApache-2.0Syntology: 21 samples, 17 ran (0 instrument-failed) · 4 unverified report
nvlabs/pac-bayes-conformal-prediction mentioned on GitHubpytorchSyntology: 3 samples, 1 ran (0 instrument-failed) · 2 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

24 samples harvested: 18 ran (of which 0 constructed an object rather than computing a result; 18 with no instrument failure: 1 honoured, 0 violated, 17 with no contract checked; 0 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
17ran
6unverified

Licence: 3 of the 24 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 2 repositories linked to this paper, official or community; the list below is grouped by repository, each group headed by its name and what it is to the paper. “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.

deepmind/conformal_training official repository 21 samples, 17 ran (0 instrument-failed) · 4 unverified
augment_flip_crop data.py ran Apache-2.0 (permissive) · 81003c641202bd0f · report fixture: signature batch:dict/0/?, shape:tuple/1/int, crop:int/0/int, mode:str/0/str, replace:int/0/int · returned dict · level recorded 2026-09-24
backward_step variational_sorting_net.py ran Apache-2.0 (permissive) · 9cad00fe020fc4a7 · report fixture: signature u:ArrayImpl/2/float, stage_idx:int/0/int, comms:dict/0/?, lam:ArrayImpl/1/float · returned ArrayImpl · level recorded 2026-09-24
blend in 12 papers auto_augment.py ran Apache-2.0 (permissive) · 985152c5e3f36838 · report fixture: signature image1:EagerTensor/3/int, image2:EagerTensor/3/int, factor:float/0/float · returned SymbolicTensor · level recorded 2026-09-24
butterfly variational_sorting_net.py ran Apache-2.0 (permissive) · 7f91d2d665604222 · report fixture: signature lam:float/0/float, x1:Tensor/2/float, x2:Tensor/2/float · returned tuple · level recorded 2026-09-24
calibrate_threshold conformal_prediction.py ran Apache-2.0 (permissive) · e807c30774b5a051 · report fixture: signature probabilities:ArrayImpl/2/float, labels:ArrayImpl/1/int, alpha:float/0/float, quantile_fn:callable/0/? · returned ArrayImpl · level recorded 2026-09-24
comm_pattern_bitonic sorting_nets.py ran fingerprinted Apache-2.0 (permissive) · 0bc4c884c3aac9a3 · report fixture: signature num_bits:int/0/int · returned dict · level recorded 2026-09-24
comm_pattern_from_list sorting_nets.py ran Apache-2.0 (permissive) · 3ac1a2c338105bb8 · report fixture: signature snet_list:list/1/list, make_parallel:bool/0/bool · returned dict · level recorded 2026-09-24
compute_accuracy in 81 papers evaluation.py ran Apache-2.0 (permissive) · 2504d3a1635ed126 · report fixture: signature probabilities:ArrayImpl/2/float, labels:ArrayImpl/1/int · returned ArrayImpl · level recorded 2026-09-24
compute_conditional_accuracy_with_checks evaluation.py ran Apache-2.0 (permissive) · dc44afe5b06af967 · report fixture: signature probabilities:ArrayImpl/2/float, labels:ArrayImpl/1/int, conditional_labels:ArrayImpl/1/int, conditional_label:int/0/int · returned ArrayImpl · level recorded 2026-09-24
conformal_quantile_with_checks conformal_prediction.py ran Apache-2.0 (permissive) · 3ce3ac6b8e6636f3 · report fixture: signature array:ArrayImpl/1/float, q:float/0/float · returned ArrayImpl · level recorded 2026-09-24
create_cnn models.py ran Apache-2.0 (permissive) · ef07ae80f9440fde · report fixture: signature classes:int/0/int, activation:str/0/str, channels:list/1/int, kernels:list/1/int, whitening:tuple/1/ArrayImpl · returned TransformedWithState · level recorded 2026-09-24
create_mlp in 17 papers models.py ran Apache-2.0 (permissive) · 78e646ce84523a39 · report fixture: signature classes:int/0/int, activation:str/0/str, units:list/1/int, whitening:tuple/1/ArrayImpl · returned TransformedWithState · level recorded 2026-09-24
forward_step variational_sorting_net.py ran Apache-2.0 (permissive) · 66379d1592d6ff49 · report fixture: signature x:ArrayImpl/1/float, stage_idx:int/0/int, comms:dict/0/?, dispersion:float/0/float, swap_prob_fun:callable/0/?, hard_swap_prob_fun:callable/0/?, key:NoneType/0/? · returned tuple · level recorded 2026-09-24
get_groups in 12 papers colab_utils.py ran Apache-2.0 (permissive) · 678c8c5bb71dae99 · report fixture: signature dataset:str/0/str, key:str/0/str · returned ArrayImpl · level recorded 2026-09-24
prune in 17 papers sorting_nets.py ran Apache-2.0 (permissive) · f0772d96f45ed98e · report fixture: signature snet_list:list/1/list, keep:list/1/int · returned list · level recorded 2026-09-24
smooth_predict_threshold smooth_conformal_prediction.py ran Apache-2.0 (permissive) · d0417a540db4e413 · report fixture: signature probabilities:ArrayImpl/2/float, tau:float/0/float, temperature:float/0/float · returned ArrayImpl · level recorded 2026-09-24
solarize in 15 papers auto_augment.py ran Apache-2.0 (permissive) · f483cc864d3d3300 · report fixture: signature image:EagerTensor/3/float, threshold:int/0/int · returned SymbolicTensor · level recorded 2026-09-24
compute_conditional_accuracy evaluation.py unverified Apache-2.0 (permissive) · 42ff9c6b22dfac9f · report
conformal_quantile in 5 papers conformal_prediction.py unverified Apache-2.0 (permissive) · 644a7a98b9d52da2 · report
create_resnet models.py unverified Apache-2.0 (permissive) · beb3dd8196db1da5 · report
cutout in 17 papers auto_augment.py unverified Apache-2.0 (permissive) · 4515269703640fc7 · report
nvlabs/pac-bayes-conformal-prediction community (archive-listed) 3 samples, 1 ran (0 instrument-failed) · 2 unverified
avg_loss_constraint src/confpred/optim.py ran · honoured contract licence not identified · pointer only · 2dc41514d7c8bde7 · report fixture: signature loss_and_constraint_func:callable/0/?, dataloader:list/1/tuple, device:str/0/str · returned tuple · level recorded 2026-09-22
aug_lag_constrained_opt src/confpred/optim.py unverified licence not identified · pointer only · feaee25fb32d4501 · report
unconstrained_opt src/confpred/optim.py unverified licence not identified · pointer only · 287d6c7e0ca58d08 · 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. For a generated assembly, an independent peer is a second implementation generated separately from the same paper text; neither implementation read the authors' code. At level 2 the two agreed numerically on the fixture's inputs; level 3 is recorded when property tests derived from the paper also passed on both, and each level 3 card prints how many its report lists. 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. A generated assembly's label was set when its code was generated for the method, and no model judged it; its card prints it as "by construction" or "not judged by a model". Where an assembly's phrase has no constructed figure, a brick that ran is not among the harvested samples this build reads, and whether a brick's run only constructed an object is read from that brick's harvested sample, so the figure was not counted; it does not say there were none.

2 assemblies on this paper: 1 harvested, 1 generated.

generated Assembly for Conformal Training (ConfTr) (graph method: Conformal Training (ConfTr)), generated from the paper text, not the repository's code. 1 brick; individually on a fixture: 1 ran (1 with no instrument failure: 1 honoured, 0 violated, 0 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified. Ran end-to-end against an independent peer (level 3; 5 property tests passed on both).

label matches by construction: the code was generated for this method (not judged by a model)

  1. smooth_quantile_conformal_threshold entry · level 3 · ran · honoured contract · 96ae61300f90ff5c5e43c03f6a874d1d2cf1019ef22364fdd82b2d8f794c62fb

harvested Assembly for Conformal Training (ConfTr) (graph method: Conformal Training (ConfTr)). 3 bricks; individually on a fixture: 1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 1 honoured, 0 violated, 0 with no contract checked; 0 where Syntology's instrument failed) · 2 unverified. The assembly has not been run end-to-end.

  1. aug_lag_constrained_opt entry · level 0 · unverified · feaee25fb32d4501
  2. avg_loss_constraint helper · level 1 · ran · honoured contract · 2dc41514d7c8bde7
  3. unconstrained_opt helper · level 0 · unverified · 287d6c7e0ca58d08

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="2110.09192") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

Conformal PredictionMedical Diagnosis

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

Test

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