Mix-n-Match: Ensemble and Compositional Methods for Uncertainty\n Calibration in Deep Learning Article Swipe
Jize Zhang
,
Bhavya Kailkhura
,
T. Yong-Jin Han
·
YOU?
·
· 2020
· Open Access
·
· DOI: https://doi.org/10.48550/arxiv.2003.07329
YOU?
·
· 2020
· Open Access
·
· DOI: https://doi.org/10.48550/arxiv.2003.07329
This paper studies the problem of post-hoc calibration of machine learning\nclassifiers. We introduce the following desiderata for uncertainty calibration:\n(a) accuracy-preserving, (b) data-efficient, and (c) high expressive power. We\nshow that none of the existing methods satisfy all three requirements, and\ndemonstrate how Mix-n-Match calibration strategies (i.e., ensemble and\ncomposition) can help achieve remarkably better data-efficiency and expressive\npower while provably maintaining the classification accuracy of the original\nclassifier. Mix-n-Match strategies are generic in the sense that they can be\nused to improve the performance of any off-the-shelf calibrator. We also reveal\npotential issues in standard evaluation practices. Popular approaches (e.g.,\nhistogram-based expected calibration error (ECE)) may provide misleading\nresults especially in small-data regime. Therefore, we propose an alternative\ndata-efficient kernel density-based estimator for a reliable evaluation of the\ncalibration performance and prove its asymptotically unbiasedness and\nconsistency. Our approaches outperform state-of-the-art solutions on both the\ncalibration as well as the evaluation tasks in most of the experimental\nsettings. Our codes are available at\nhttps://github.com/zhang64-llnl/Mix-n-Match-Calibration.\n
Related Topics
Concepts
Metadata
- Type
- preprint
- Landing Page
- http://arxiv.org/abs/2003.07329
- https://arxiv.org/pdf/2003.07329
- OA Status
- green
- Cited By
- 47
- Related Works
- 10
- OpenAlex ID
- https://openalex.org/W3034905393
All OpenAlex metadata
Raw OpenAlex JSON
- OpenAlex ID
-
https://openalex.org/W3034905393Canonical identifier for this work in OpenAlex
- DOI
-
https://doi.org/10.48550/arxiv.2003.07329Digital Object Identifier
- Title
-
Mix-n-Match: Ensemble and Compositional Methods for Uncertainty\n Calibration in Deep LearningWork title
- Type
-
preprintOpenAlex work type
- Publication year
-
2020Year of publication
- Publication date
-
2020-03-16Full publication date if available
- Authors
-
Jize Zhang, Bhavya Kailkhura, T. Yong-Jin HanList of authors in order
- Landing page
-
https://arxiv.org/abs/2003.07329Publisher landing page
- PDF URL
-
https://arxiv.org/pdf/2003.07329Direct link to full text PDF
- Open access
-
YesWhether a free full text is available
- OA status
-
greenOpen access status per OpenAlex
- OA URL
-
https://arxiv.org/pdf/2003.07329Direct OA link when available
- Concepts
-
Computer science, Calibration, Estimator, Classifier (UML), Consistency (knowledge bases), Ensemble learning, Artificial intelligence, Machine learning, Kernel (algebra), Data mining, Algorithm, Statistics, Mathematics, CombinatoricsTop concepts (fields/topics) attached by OpenAlex
- Cited by
-
47Total citation count in OpenAlex
- Citations by year (recent)
-
2024: 5, 2023: 8, 2022: 10, 2021: 15, 2020: 9Per-year citation counts (last 5 years)
- Related works (count)
-
10Other works algorithmically related by OpenAlex
Full payload
| id | https://openalex.org/W3034905393 |
|---|---|
| doi | https://doi.org/10.48550/arxiv.2003.07329 |
| ids.mag | 3034905393 |
| ids.openalex | https://openalex.org/W3034905393 |
| fwci | 6.16810104 |
| type | preprint |
| title | Mix-n-Match: Ensemble and Compositional Methods for Uncertainty\n Calibration in Deep Learning |
| biblio.issue | |
| biblio.volume | |
| biblio.last_page | |
| biblio.first_page | |
| topics[0].id | https://openalex.org/T11512 |
| topics[0].field.id | https://openalex.org/fields/17 |
| topics[0].field.display_name | Computer Science |
| topics[0].score | 0.9994000196456909 |
| topics[0].domain.id | https://openalex.org/domains/3 |
| topics[0].domain.display_name | Physical Sciences |
| topics[0].subfield.id | https://openalex.org/subfields/1702 |
| topics[0].subfield.display_name | Artificial Intelligence |
| topics[0].display_name | Anomaly Detection Techniques and Applications |
| topics[1].id | https://openalex.org/T11689 |
| topics[1].field.id | https://openalex.org/fields/17 |
| topics[1].field.display_name | Computer Science |
| topics[1].score | 0.9976999759674072 |
| topics[1].domain.id | https://openalex.org/domains/3 |
| topics[1].domain.display_name | Physical Sciences |
| topics[1].subfield.id | https://openalex.org/subfields/1702 |
| topics[1].subfield.display_name | Artificial Intelligence |
| topics[1].display_name | Adversarial Robustness in Machine Learning |
| topics[2].id | https://openalex.org/T12535 |
| topics[2].field.id | https://openalex.org/fields/17 |
| topics[2].field.display_name | Computer Science |
| topics[2].score | 0.9965000152587891 |
| topics[2].domain.id | https://openalex.org/domains/3 |
| topics[2].domain.display_name | Physical Sciences |
| topics[2].subfield.id | https://openalex.org/subfields/1702 |
| topics[2].subfield.display_name | Artificial Intelligence |
| topics[2].display_name | Machine Learning and Data Classification |
| is_xpac | False |
| apc_list | |
| apc_paid | |
| concepts[0].id | https://openalex.org/C41008148 |
| concepts[0].level | 0 |
| concepts[0].score | 0.7500623464584351 |
| concepts[0].wikidata | https://www.wikidata.org/wiki/Q21198 |
| concepts[0].display_name | Computer science |
| concepts[1].id | https://openalex.org/C165838908 |
| concepts[1].level | 2 |
| concepts[1].score | 0.7092698812484741 |
| concepts[1].wikidata | https://www.wikidata.org/wiki/Q736777 |
| concepts[1].display_name | Calibration |
| concepts[2].id | https://openalex.org/C185429906 |
| concepts[2].level | 2 |
| concepts[2].score | 0.6577190160751343 |
| concepts[2].wikidata | https://www.wikidata.org/wiki/Q1130160 |
| concepts[2].display_name | Estimator |
| concepts[3].id | https://openalex.org/C95623464 |
| concepts[3].level | 2 |
| concepts[3].score | 0.6523599624633789 |
| concepts[3].wikidata | https://www.wikidata.org/wiki/Q1096149 |
| concepts[3].display_name | Classifier (UML) |
| concepts[4].id | https://openalex.org/C2776436953 |
| concepts[4].level | 2 |
| concepts[4].score | 0.5897754430770874 |
| concepts[4].wikidata | https://www.wikidata.org/wiki/Q5163215 |
| concepts[4].display_name | Consistency (knowledge bases) |
| concepts[5].id | https://openalex.org/C45942800 |
| concepts[5].level | 2 |
| concepts[5].score | 0.538857638835907 |
| concepts[5].wikidata | https://www.wikidata.org/wiki/Q245652 |
| concepts[5].display_name | Ensemble learning |
| concepts[6].id | https://openalex.org/C154945302 |
| concepts[6].level | 1 |
| concepts[6].score | 0.5321570634841919 |
| concepts[6].wikidata | https://www.wikidata.org/wiki/Q11660 |
| concepts[6].display_name | Artificial intelligence |
| concepts[7].id | https://openalex.org/C119857082 |
| concepts[7].level | 1 |
| concepts[7].score | 0.5203644633293152 |
| concepts[7].wikidata | https://www.wikidata.org/wiki/Q2539 |
| concepts[7].display_name | Machine learning |
| concepts[8].id | https://openalex.org/C74193536 |
| concepts[8].level | 2 |
| concepts[8].score | 0.5197789072990417 |
| concepts[8].wikidata | https://www.wikidata.org/wiki/Q574844 |
| concepts[8].display_name | Kernel (algebra) |
| concepts[9].id | https://openalex.org/C124101348 |
| concepts[9].level | 1 |
| concepts[9].score | 0.3806314766407013 |
| concepts[9].wikidata | https://www.wikidata.org/wiki/Q172491 |
| concepts[9].display_name | Data mining |
| concepts[10].id | https://openalex.org/C11413529 |
| concepts[10].level | 1 |
| concepts[10].score | 0.34736132621765137 |
| concepts[10].wikidata | https://www.wikidata.org/wiki/Q8366 |
| concepts[10].display_name | Algorithm |
| concepts[11].id | https://openalex.org/C105795698 |
| concepts[11].level | 1 |
| concepts[11].score | 0.17986822128295898 |
| concepts[11].wikidata | https://www.wikidata.org/wiki/Q12483 |
| concepts[11].display_name | Statistics |
| concepts[12].id | https://openalex.org/C33923547 |
| concepts[12].level | 0 |
| concepts[12].score | 0.17941215634346008 |
| concepts[12].wikidata | https://www.wikidata.org/wiki/Q395 |
| concepts[12].display_name | Mathematics |
| concepts[13].id | https://openalex.org/C114614502 |
| concepts[13].level | 1 |
| concepts[13].score | 0.0 |
| concepts[13].wikidata | https://www.wikidata.org/wiki/Q76592 |
| concepts[13].display_name | Combinatorics |
| keywords[0].id | https://openalex.org/keywords/computer-science |
| keywords[0].score | 0.7500623464584351 |
| keywords[0].display_name | Computer science |
| keywords[1].id | https://openalex.org/keywords/calibration |
| keywords[1].score | 0.7092698812484741 |
| keywords[1].display_name | Calibration |
| keywords[2].id | https://openalex.org/keywords/estimator |
| keywords[2].score | 0.6577190160751343 |
| keywords[2].display_name | Estimator |
| keywords[3].id | https://openalex.org/keywords/classifier |
| keywords[3].score | 0.6523599624633789 |
| keywords[3].display_name | Classifier (UML) |
| keywords[4].id | https://openalex.org/keywords/consistency |
| keywords[4].score | 0.5897754430770874 |
| keywords[4].display_name | Consistency (knowledge bases) |
| keywords[5].id | https://openalex.org/keywords/ensemble-learning |
| keywords[5].score | 0.538857638835907 |
| keywords[5].display_name | Ensemble learning |
| keywords[6].id | https://openalex.org/keywords/artificial-intelligence |
| keywords[6].score | 0.5321570634841919 |
| keywords[6].display_name | Artificial intelligence |
| keywords[7].id | https://openalex.org/keywords/machine-learning |
| keywords[7].score | 0.5203644633293152 |
| keywords[7].display_name | Machine learning |
| keywords[8].id | https://openalex.org/keywords/kernel |
| keywords[8].score | 0.5197789072990417 |
| keywords[8].display_name | Kernel (algebra) |
| keywords[9].id | https://openalex.org/keywords/data-mining |
| keywords[9].score | 0.3806314766407013 |
| keywords[9].display_name | Data mining |
| keywords[10].id | https://openalex.org/keywords/algorithm |
| keywords[10].score | 0.34736132621765137 |
| keywords[10].display_name | Algorithm |
| keywords[11].id | https://openalex.org/keywords/statistics |
| keywords[11].score | 0.17986822128295898 |
| keywords[11].display_name | Statistics |
| keywords[12].id | https://openalex.org/keywords/mathematics |
| keywords[12].score | 0.17941215634346008 |
| keywords[12].display_name | Mathematics |
| language | |
| locations[0].id | pmh:oai:arXiv.org:2003.07329 |
| locations[0].is_oa | True |
| locations[0].source.id | https://openalex.org/S4306400194 |
| locations[0].source.issn | |
| locations[0].source.type | repository |
| locations[0].source.is_oa | True |
| locations[0].source.issn_l | |
| locations[0].source.is_core | False |
| locations[0].source.is_in_doaj | False |
| locations[0].source.display_name | arXiv (Cornell University) |
| locations[0].source.host_organization | https://openalex.org/I205783295 |
| locations[0].source.host_organization_name | Cornell University |
| locations[0].source.host_organization_lineage | https://openalex.org/I205783295 |
| locations[0].license | |
| locations[0].pdf_url | https://arxiv.org/pdf/2003.07329 |
| locations[0].version | submittedVersion |
| locations[0].raw_type | text |
| locations[0].license_id | |
| locations[0].is_accepted | False |
| locations[0].is_published | False |
| locations[0].raw_source_name | |
| locations[0].landing_page_url | http://arxiv.org/abs/2003.07329 |
| indexed_in | arxiv |
| authorships[0].author.id | https://openalex.org/A5060905917 |
| authorships[0].author.orcid | https://orcid.org/0000-0002-0553-3979 |
| authorships[0].author.display_name | Jize Zhang |
| authorships[0].author_position | first |
| authorships[0].raw_author_name | Jize Zhang |
| authorships[0].is_corresponding | False |
| authorships[1].author.id | https://openalex.org/A5041470575 |
| authorships[1].author.orcid | https://orcid.org/0000-0002-2819-2919 |
| authorships[1].author.display_name | Bhavya Kailkhura |
| authorships[1].author_position | middle |
| authorships[1].raw_author_name | Bhavya Kailkhura |
| authorships[1].is_corresponding | False |
| authorships[2].author.id | https://openalex.org/A5067913179 |
| authorships[2].author.orcid | https://orcid.org/0000-0002-3000-2782 |
| authorships[2].author.display_name | T. Yong-Jin Han |
| authorships[2].author_position | last |
| authorships[2].raw_author_name | T. Yong-Jin Han |
| authorships[2].is_corresponding | False |
| has_content.pdf | False |
| has_content.grobid_xml | False |
| is_paratext | False |
| open_access.is_oa | True |
| open_access.oa_url | https://arxiv.org/pdf/2003.07329 |
| open_access.oa_status | green |
| open_access.any_repository_has_fulltext | False |
| created_date | 2020-06-19T00:00:00 |
| display_name | Mix-n-Match: Ensemble and Compositional Methods for Uncertainty\n Calibration in Deep Learning |
| has_fulltext | False |
| is_retracted | False |
| updated_date | 2025-11-06T03:46:38.306776 |
| primary_topic.id | https://openalex.org/T11512 |
| primary_topic.field.id | https://openalex.org/fields/17 |
| primary_topic.field.display_name | Computer Science |
| primary_topic.score | 0.9994000196456909 |
| primary_topic.domain.id | https://openalex.org/domains/3 |
| primary_topic.domain.display_name | Physical Sciences |
| primary_topic.subfield.id | https://openalex.org/subfields/1702 |
| primary_topic.subfield.display_name | Artificial Intelligence |
| primary_topic.display_name | Anomaly Detection Techniques and Applications |
| related_works | https://openalex.org/W4287880334, https://openalex.org/W4366700029, https://openalex.org/W4285230481, https://openalex.org/W4385769873, https://openalex.org/W2015759683, https://openalex.org/W4281634296, https://openalex.org/W4319161863, https://openalex.org/W2371687270, https://openalex.org/W4307819175, https://openalex.org/W4311888330 |
| cited_by_count | 47 |
| counts_by_year[0].year | 2024 |
| counts_by_year[0].cited_by_count | 5 |
| counts_by_year[1].year | 2023 |
| counts_by_year[1].cited_by_count | 8 |
| counts_by_year[2].year | 2022 |
| counts_by_year[2].cited_by_count | 10 |
| counts_by_year[3].year | 2021 |
| counts_by_year[3].cited_by_count | 15 |
| counts_by_year[4].year | 2020 |
| counts_by_year[4].cited_by_count | 9 |
| locations_count | 1 |
| best_oa_location.id | pmh:oai:arXiv.org:2003.07329 |
| best_oa_location.is_oa | True |
| best_oa_location.source.id | https://openalex.org/S4306400194 |
| best_oa_location.source.issn | |
| best_oa_location.source.type | repository |
| best_oa_location.source.is_oa | True |
| best_oa_location.source.issn_l | |
| best_oa_location.source.is_core | False |
| best_oa_location.source.is_in_doaj | False |
| best_oa_location.source.display_name | arXiv (Cornell University) |
| best_oa_location.source.host_organization | https://openalex.org/I205783295 |
| best_oa_location.source.host_organization_name | Cornell University |
| best_oa_location.source.host_organization_lineage | https://openalex.org/I205783295 |
| best_oa_location.license | |
| best_oa_location.pdf_url | https://arxiv.org/pdf/2003.07329 |
| best_oa_location.version | submittedVersion |
| best_oa_location.raw_type | text |
| best_oa_location.license_id | |
| best_oa_location.is_accepted | False |
| best_oa_location.is_published | False |
| best_oa_location.raw_source_name | |
| best_oa_location.landing_page_url | http://arxiv.org/abs/2003.07329 |
| primary_location.id | pmh:oai:arXiv.org:2003.07329 |
| primary_location.is_oa | True |
| primary_location.source.id | https://openalex.org/S4306400194 |
| primary_location.source.issn | |
| primary_location.source.type | repository |
| primary_location.source.is_oa | True |
| primary_location.source.issn_l | |
| primary_location.source.is_core | False |
| primary_location.source.is_in_doaj | False |
| primary_location.source.display_name | arXiv (Cornell University) |
| primary_location.source.host_organization | https://openalex.org/I205783295 |
| primary_location.source.host_organization_name | Cornell University |
| primary_location.source.host_organization_lineage | https://openalex.org/I205783295 |
| primary_location.license | |
| primary_location.pdf_url | https://arxiv.org/pdf/2003.07329 |
| primary_location.version | submittedVersion |
| primary_location.raw_type | text |
| primary_location.license_id | |
| primary_location.is_accepted | False |
| primary_location.is_published | False |
| primary_location.raw_source_name | |
| primary_location.landing_page_url | http://arxiv.org/abs/2003.07329 |
| publication_date | 2020-03-16 |
| publication_year | 2020 |
| referenced_works_count | 0 |
| abstract_inverted_index.a | 113 |
| abstract_inverted_index.We | 11, 82 |
| abstract_inverted_index.an | 107 |
| abstract_inverted_index.as | 133, 135 |
| abstract_inverted_index.in | 67, 86, 101, 139 |
| abstract_inverted_index.of | 5, 8, 30, 60, 78, 116, 141 |
| abstract_inverted_index.on | 130 |
| abstract_inverted_index.to | 74 |
| abstract_inverted_index.we | 105 |
| abstract_inverted_index.(b) | 20 |
| abstract_inverted_index.(c) | 23 |
| abstract_inverted_index.Our | 125, 144 |
| abstract_inverted_index.all | 35 |
| abstract_inverted_index.and | 22, 52, 119 |
| abstract_inverted_index.any | 79 |
| abstract_inverted_index.are | 65, 146 |
| abstract_inverted_index.can | 46, 72 |
| abstract_inverted_index.for | 16, 112 |
| abstract_inverted_index.how | 39 |
| abstract_inverted_index.its | 121 |
| abstract_inverted_index.may | 97 |
| abstract_inverted_index.the | 3, 13, 31, 57, 61, 68, 76, 136, 142 |
| abstract_inverted_index.This | 0 |
| abstract_inverted_index.also | 83 |
| abstract_inverted_index.both | 131 |
| abstract_inverted_index.help | 47 |
| abstract_inverted_index.high | 24 |
| abstract_inverted_index.most | 140 |
| abstract_inverted_index.none | 29 |
| abstract_inverted_index.that | 28, 70 |
| abstract_inverted_index.they | 71 |
| abstract_inverted_index.well | 134 |
| abstract_inverted_index.codes | 145 |
| abstract_inverted_index.error | 95 |
| abstract_inverted_index.paper | 1 |
| abstract_inverted_index.prove | 120 |
| abstract_inverted_index.sense | 69 |
| abstract_inverted_index.tasks | 138 |
| abstract_inverted_index.three | 36 |
| abstract_inverted_index.while | 54 |
| abstract_inverted_index.(ECE)) | 96 |
| abstract_inverted_index.(i.e., | 43 |
| abstract_inverted_index.better | 50 |
| abstract_inverted_index.issues | 85 |
| abstract_inverted_index.kernel | 109 |
| abstract_inverted_index.power. | 26 |
| abstract_inverted_index.Popular | 90 |
| abstract_inverted_index.achieve | 48 |
| abstract_inverted_index.generic | 66 |
| abstract_inverted_index.improve | 75 |
| abstract_inverted_index.machine | 9 |
| abstract_inverted_index.methods | 33 |
| abstract_inverted_index.problem | 4 |
| abstract_inverted_index.propose | 106 |
| abstract_inverted_index.provide | 98 |
| abstract_inverted_index.regime. | 103 |
| abstract_inverted_index.satisfy | 34 |
| abstract_inverted_index.studies | 2 |
| abstract_inverted_index.We\nshow | 27 |
| abstract_inverted_index.accuracy | 59 |
| abstract_inverted_index.be\nused | 73 |
| abstract_inverted_index.ensemble | 44 |
| abstract_inverted_index.existing | 32 |
| abstract_inverted_index.expected | 93 |
| abstract_inverted_index.post-hoc | 6 |
| abstract_inverted_index.provably | 55 |
| abstract_inverted_index.reliable | 114 |
| abstract_inverted_index.standard | 87 |
| abstract_inverted_index.available | 147 |
| abstract_inverted_index.estimator | 111 |
| abstract_inverted_index.following | 14 |
| abstract_inverted_index.introduce | 12 |
| abstract_inverted_index.solutions | 129 |
| abstract_inverted_index.Therefore, | 104 |
| abstract_inverted_index.approaches | 91, 126 |
| abstract_inverted_index.desiderata | 15 |
| abstract_inverted_index.especially | 100 |
| abstract_inverted_index.evaluation | 88, 115, 137 |
| abstract_inverted_index.expressive | 25 |
| abstract_inverted_index.outperform | 127 |
| abstract_inverted_index.practices. | 89 |
| abstract_inverted_index.remarkably | 49 |
| abstract_inverted_index.small-data | 102 |
| abstract_inverted_index.strategies | 42, 64 |
| abstract_inverted_index.Mix-n-Match | 40, 63 |
| abstract_inverted_index.calibration | 7, 41, 94 |
| abstract_inverted_index.calibrator. | 81 |
| abstract_inverted_index.maintaining | 56 |
| abstract_inverted_index.performance | 77, 118 |
| abstract_inverted_index.uncertainty | 17 |
| abstract_inverted_index.unbiasedness | 123 |
| abstract_inverted_index.density-based | 110 |
| abstract_inverted_index.off-the-shelf | 80 |
| abstract_inverted_index.requirements, | 37 |
| abstract_inverted_index.asymptotically | 122 |
| abstract_inverted_index.classification | 58 |
| abstract_inverted_index.data-efficiency | 51 |
| abstract_inverted_index.data-efficient, | 21 |
| abstract_inverted_index.and\ndemonstrate | 38 |
| abstract_inverted_index.state-of-the-art | 128 |
| abstract_inverted_index.the\ncalibration | 117, 132 |
| abstract_inverted_index.and\ncomposition) | 45 |
| abstract_inverted_index.and\nconsistency. | 124 |
| abstract_inverted_index.calibration:\n(a) | 18 |
| abstract_inverted_index.expressive\npower | 53 |
| abstract_inverted_index.reveal\npotential | 84 |
| abstract_inverted_index.misleading\nresults | 99 |
| abstract_inverted_index.accuracy-preserving, | 19 |
| abstract_inverted_index.original\nclassifier. | 62 |
| abstract_inverted_index.learning\nclassifiers. | 10 |
| abstract_inverted_index.(e.g.,\nhistogram-based | 92 |
| abstract_inverted_index.experimental\nsettings. | 143 |
| abstract_inverted_index.alternative\ndata-efficient | 108 |
| abstract_inverted_index.at\nhttps://github.com/zhang64-llnl/Mix-n-Match-Calibration.\n | 148 |
| cited_by_percentile_year.max | 99 |
| cited_by_percentile_year.min | 98 |
| countries_distinct_count | 0 |
| institutions_distinct_count | 3 |
| citation_normalized_percentile.value | 0.96722072 |
| citation_normalized_percentile.is_in_top_1_percent | False |
| citation_normalized_percentile.is_in_top_10_percent | True |