Search results

Create a live benchmark / leaderboard mini website for PU and PUN problems

area:benchmarksBenchmark harness and datasetsBenchmark harness and datasets

documentationImprovements or additions to documentationImprovements or additions to documentation

effort:XLExtra large effortExtra large effort

enhancementNew feature or requestNew feature or request

priority:P2Lower priorityLower priority

type:benchmarkBenchmarking workBenchmarking work

Status: Open.

#213 In pulearn/pulearn;

· shaypal5 opened on Apr 11, 2026

Bug: BaggingPuClassifier incompatible with scikit-learn ≥ 1.6 — partial_dependence and is_classifier fail

Status: Open.

#176 In pulearn/pulearn;

· MaxEtherington opened on Mar 12, 2026

Tighten PriorEstimateResult confidence_interval typing

area:priorsClass-prior estimationClass-prior estimation

effort:SSmall effortSmall effort

priority:P2Lower priorityLower priority

type:refactorRefactor workRefactor work

Status: Open.

#162 In pulearn/pulearn;

· shaypal5 opened on Mar 6, 2026

Backwards compatibility policy + deprecation pathway (labels, API changes, release notes template)

area:apiPublic API and contractsPublic API and contracts

effort:SSmall effortSmall effort

priority:P1Medium-high priorityMedium-high priority

status:blockedBlocked by dependenciesBlocked by dependencies

type:infraInfrastructure workInfrastructure work

Status: Open.

#152 In pulearn/pulearn;

· shaypal5 opened on Mar 5, 2026

Sample-weight behavior tests and docs

area:qualityQuality, interop, and release readinessQuality, interop, and release readiness

effort:MMedium effortMedium effort

priority:P1Medium-high priorityMedium-high priority

status:blockedBlocked by dependenciesBlocked by dependencies

type:testsTesting workTesting work

Status: Open.

#151 In pulearn/pulearn;

· shaypal5 opened on Mar 5, 2026

Sparse matrix + large n tests (smoke)

area:qualityQuality, interop, and release readinessQuality, interop, and release readiness

effort:MMedium effortMedium effort

priority:P1Medium-high priorityMedium-high priority

status:blockedBlocked by dependenciesBlocked by dependencies

type:testsTesting workTesting work

Status: Open.

#150 In pulearn/pulearn;

· shaypal5 opened on Mar 5, 2026

Pipeline compatibility tests for all estimators

area:qualityQuality, interop, and release readinessQuality, interop, and release readiness

effort:MMedium effortMedium effort

priority:P1Medium-high priorityMedium-high priority

status:blockedBlocked by dependenciesBlocked by dependencies

type:testsTesting workTesting work

Status: Open.

#149 In pulearn/pulearn;

· shaypal5 opened on Mar 5, 2026

Interoperability: sklearn pipelines, ColumnTransformer, sparse matrices, sample weights, multiprocessing stability

area:qualityQuality, interop, and release readinessQuality, interop, and release readiness

effort:LLarge effortLarge effort

priority:P1Medium-high priorityMedium-high priority

status:blockedBlocked by dependenciesBlocked by dependencies

type:testsTesting workTesting work

Status: Open.

#148 In pulearn/pulearn;

· shaypal5 opened on Mar 5, 2026

Examples overhaul: end-to-end notebooks/scripts for prior estimation + training + evaluation + benchmarking

area:docsDocumentation areaDocumentation area

effort:MMedium effortMedium effort

priority:P1Medium-high priorityMedium-high priority

status:blockedBlocked by dependenciesBlocked by dependencies

type:docsDocumentation workDocumentation work

Status: Open.

#147 In pulearn/pulearn;

· shaypal5 opened on Mar 5, 2026

“Failure modes playbook”: selection bias, distribution shift, leakage, degenerate predictors; recommended mitigations

area:docsDocumentation areaDocumentation area

assumption:SARAssumes SAR labelingAssumes SAR labeling

assumption:SCARAssumes SCAR labelingAssumes SCAR labeling

effort:MMedium effortMedium effort

priority:P0Highest priorityHighest priority

status:blockedBlocked by dependenciesBlocked by dependencies

type:docsDocumentation workDocumentation work

Status: Open.

#146 In pulearn/pulearn;

· shaypal5 opened on Mar 5, 2026

“How to evaluate”: why standard metrics break; corrected metrics; when π/c are required; how to estimate them

area:docsDocumentation areaDocumentation area

assumption:SCARAssumes SCAR labelingAssumes SCAR labeling

effort:MMedium effortMedium effort

priority:P0Highest priorityHighest priority

status:blockedBlocked by dependenciesBlocked by dependencies

type:docsDocumentation workDocumentation work

Status: Open.

#145 In pulearn/pulearn;

· shaypal5 opened on Mar 5, 2026

“How to choose a learner”: guidance matrix (data size, model class, stability, calibration)

area:docsDocumentation areaDocumentation area

assumption:SCARAssumes SCAR labelingAssumes SCAR labeling

effort:MMedium effortMedium effort

priority:P0Highest priorityHighest priority

status:blockedBlocked by dependenciesBlocked by dependencies

type:docsDocumentation workDocumentation work

Status: Open.

#144 In pulearn/pulearn;

· shaypal5 opened on Mar 5, 2026