{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,10,12]],"date-time":"2025-10-12T03:20:40Z","timestamp":1760239240657,"version":"build-2065373602"},"reference-count":52,"publisher":"MDPI AG","issue":"4","license":[{"start":{"date-parts":[[2020,10,30]],"date-time":"2020-10-30T00:00:00Z","timestamp":1604016000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100008530","name":"European Regional Development Fund","doi-asserted-by":"publisher","award":["POCI-01-0247-FEDER-033479"],"award-info":[{"award-number":["POCI-01-0247-FEDER-033479"]}],"id":[{"id":"10.13039\/501100008530","id-type":"DOI","asserted-by":"publisher"}]},{"name":"Operational Programme Competitiveness and Internationalization","award":["POCI-01-0247-FEDER-033479"],"award-info":[{"award-number":["POCI-01-0247-FEDER-033479"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["MAKE"],"abstract":"<jats:p>Uncertainty is ubiquitous and happens in every single prediction of Machine Learning models. The ability to estimate and quantify the uncertainty of individual predictions is arguably relevant, all the more in safety-critical applications. Real-world recognition poses multiple challenges since a model\u2019s knowledge about physical phenomenon is not complete, and observations are incomplete by definition. However, Machine Learning algorithms often assume that train and test data distributions are the same and that all testing classes are present during training. A more realistic scenario is the Open Set Recognition, where unknown classes can be submitted to an algorithm during testing. In this paper, we propose a Knowledge Uncertainty Estimation (KUE) method to quantify knowledge uncertainty and reject out-of-distribution inputs. Additionally, we quantify and distinguish aleatoric and epistemic uncertainty with the classical information-theoretical measures of entropy by means of ensemble techniques. We performed experiments on four datasets with different data modalities and compared our results with distance-based classifiers, SVM-based approaches and ensemble techniques using entropy measures. Overall, the effectiveness of KUE in distinguishing in- and out-distribution inputs obtained better results in most cases and was at least comparable in others. Furthermore, a classification with rejection option based on a proposed combination strategy between different measures of uncertainty is an application of uncertainty with proven results.<\/jats:p>","DOI":"10.3390\/make2040028","type":"journal-article","created":{"date-parts":[[2020,10,30]],"date-time":"2020-10-30T21:34:47Z","timestamp":1604093687000},"page":"505-532","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":7,"title":["Towards Knowledge Uncertainty Estimation for Open Set Recognition"],"prefix":"10.3390","volume":"2","author":[{"given":"Catarina","family":"Pires","sequence":"first","affiliation":[{"name":"Associa\u00e7\u00e3o Fraunhofer Portugal Research, Rua Alfredo Allen 455\/461, 4200-135 Porto, Portugal"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-9445-4809","authenticated-orcid":false,"given":"Mar\u00edlia","family":"Barandas","sequence":"additional","affiliation":[{"name":"Associa\u00e7\u00e3o Fraunhofer Portugal Research, Rua Alfredo Allen 455\/461, 4200-135 Porto, Portugal"},{"name":"Laborat\u00f3rio de Instrumenta\u00e7\u00e3o, Engenharia Biom\u00e9dica e F\u00edsica da Radia\u00e7\u00e3o (LIBPhys-UNL), Departamento de F\u00edsica, Faculdade de Ci\u00eancias e Tecnologia (FCT), Universidade Nova de Lisboa, 2829-516 Caparica, Portugal"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1624-7716","authenticated-orcid":false,"given":"Let\u00edcia","family":"Fernandes","sequence":"additional","affiliation":[{"name":"Associa\u00e7\u00e3o Fraunhofer Portugal Research, Rua Alfredo Allen 455\/461, 4200-135 Porto, Portugal"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8481-6079","authenticated-orcid":false,"given":"Duarte","family":"Folgado","sequence":"additional","affiliation":[{"name":"Associa\u00e7\u00e3o Fraunhofer Portugal Research, Rua Alfredo Allen 455\/461, 4200-135 Porto, Portugal"},{"name":"Laborat\u00f3rio de Instrumenta\u00e7\u00e3o, Engenharia Biom\u00e9dica e F\u00edsica da Radia\u00e7\u00e3o (LIBPhys-UNL), Departamento de F\u00edsica, Faculdade de Ci\u00eancias e Tecnologia (FCT), Universidade Nova de Lisboa, 2829-516 Caparica, Portugal"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4022-7424","authenticated-orcid":false,"given":"Hugo","family":"Gamboa","sequence":"additional","affiliation":[{"name":"Associa\u00e7\u00e3o Fraunhofer Portugal Research, Rua Alfredo Allen 455\/461, 4200-135 Porto, Portugal"},{"name":"Laborat\u00f3rio de Instrumenta\u00e7\u00e3o, Engenharia Biom\u00e9dica e F\u00edsica da Radia\u00e7\u00e3o (LIBPhys-UNL), Departamento de F\u00edsica, Faculdade de Ci\u00eancias e Tecnologia (FCT), Universidade Nova de Lisboa, 2829-516 Caparica, Portugal"}]}],"member":"1968","published-online":{"date-parts":[[2020,10,30]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","first-page":"20","DOI":"10.1038\/s42256-018-0004-1","article-title":"The need for uncertainty quantification in machine-assisted medical decision making","volume":"1","author":"Begoli","year":"2019","journal-title":"Nat. Mach. Intell."},{"key":"ref_2","unstructured":"H\u00fcllermeier, E., and Waegeman, W. (2019). Aleatoric and epistemic uncertainty in machine learning: A tutorial introduction. arXiv."},{"key":"ref_3","unstructured":"Schulam, P., and Saria, S. (2019). Can you trust this prediction? Auditing pointwise reliability after learning. arXiv."},{"key":"ref_4","doi-asserted-by":"crossref","first-page":"e1312","DOI":"10.1002\/widm.1312","article-title":"Causability and explainability of artificial intelligence in medicine","volume":"9","author":"Holzinger","year":"2019","journal-title":"Wiley Interdiscip. Rev. Data Min. Knowl. Discov."},{"key":"ref_5","doi-asserted-by":"crossref","unstructured":"Campagner, A., Cabitza, F., and Ciucci, D. (2020). Three-Way Decision for Handling Uncertainty in Machine Learning: A Narrative Review. International Joint Conference on Rough Sets, Springer.","DOI":"10.1007\/978-3-030-52705-1_10"},{"key":"ref_6","unstructured":"Gal, Y. (2016). Uncertainty in Deep Learning, University of Cambridge."},{"key":"ref_7","doi-asserted-by":"crossref","unstructured":"Bota, P., Silva, J., Folgado, D., and Gamboa, H. (2019). A semi-automatic annotation approach for human activity recognition. Sensors, 19.","DOI":"10.3390\/s19030501"},{"key":"ref_8","unstructured":"Mukherjee, S., and Awadallah, A.H. (2020). Uncertainty-aware Self-training for Text Classification with Few Labels. arXiv."},{"key":"ref_9","doi-asserted-by":"crossref","unstructured":"Santos, R., Barandas, M., Leonardo, R., and Gamboa, H. (2019). Fingerprints and floor plans construction for indoor localisation based on crowdsourcing. Sensors, 19.","DOI":"10.3390\/s19040919"},{"key":"ref_10","doi-asserted-by":"crossref","first-page":"2463","DOI":"10.1109\/TKDE.2012.179","article-title":"Dealing with uncertainty: A survey of theories and practices","volume":"25","author":"Li","year":"2012","journal-title":"IEEE Trans. Knowl. Data Eng."},{"key":"ref_11","doi-asserted-by":"crossref","first-page":"246","DOI":"10.1089\/big.2016.0051","article-title":"On the safety of machine learning: Cyber-physical systems, decision sciences, and data products","volume":"5","author":"Varshney","year":"2017","journal-title":"Big Data"},{"key":"ref_12","doi-asserted-by":"crossref","first-page":"16","DOI":"10.1016\/j.ins.2013.07.030","article-title":"Reliable classification: Learning classifiers that distinguish aleatoric and epistemic uncertainty","volume":"255","author":"Senge","year":"2014","journal-title":"Inf. Sci."},{"key":"ref_13","doi-asserted-by":"crossref","first-page":"1757","DOI":"10.1109\/TPAMI.2012.256","article-title":"Toward open set recognition","volume":"35","author":"Scheirer","year":"2012","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"ref_14","unstructured":"Boult, T.E., Cruz, S., Dhamija, A.R., Gunther, M., Henrydoss, J., and Scheirer, W.J. (2020, January 7\u201312). Learning and the unknown: Surveying steps toward open world recognition. Proceedings of the AAAI Conference on Artificial Intelligence, New York, NY, USA."},{"key":"ref_15","unstructured":"Ren, J., Liu, P.J., Fertig, E., Snoek, J., Poplin, R., Depristo, M., Dillon, J., and Lakshminarayanan, B. (2019). Likelihood ratios for out-of-distribution detection. Advances in Neural Information Processing Systems, MIT Press."},{"key":"ref_16","doi-asserted-by":"crossref","first-page":"2317","DOI":"10.1109\/TPAMI.2014.2321392","article-title":"Probability models for open set recognition","volume":"36","author":"Scheirer","year":"2014","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"ref_17","doi-asserted-by":"crossref","first-page":"41","DOI":"10.1109\/TIT.1970.1054406","article-title":"On optimum recognition error and reject tradeoff","volume":"16","author":"Chow","year":"1970","journal-title":"IEEE Trans. Inf. Theory"},{"key":"ref_18","doi-asserted-by":"crossref","unstructured":"Zhu, L., and Laptev, N. (2017, January 18\u201321). Deep and confident prediction for time series at uber. Proceedings of the 2017 IEEE International Conference on Data Mining Workshops (ICDMW), New Orleans, LA, USA.","DOI":"10.1109\/ICDMW.2017.19"},{"key":"ref_19","unstructured":"Malinin, A., and Gales, M. (2018). Predictive uncertainty estimation via prior networks. Advances in Neural Information Processing Systems, Curran Associates, Inc."},{"key":"ref_20","doi-asserted-by":"crossref","first-page":"1565","DOI":"10.1016\/j.patrec.2008.03.010","article-title":"Growing a multi-class classifier with a reject option","volume":"29","author":"Tax","year":"2008","journal-title":"Pattern Recognit. Lett."},{"key":"ref_21","doi-asserted-by":"crossref","first-page":"2099","DOI":"10.1016\/S0031-3203(00)00059-5","article-title":"Reject option with multiple thresholds","volume":"33","author":"Fumera","year":"2000","journal-title":"Pattern Recognit."},{"key":"ref_22","doi-asserted-by":"crossref","first-page":"361","DOI":"10.1109\/3468.487961","article-title":"Representing partial ignorance","volume":"26","author":"Dubois","year":"1996","journal-title":"IEEE Trans. Syst. Man Cybern. Part A Syst. Hum."},{"key":"ref_23","doi-asserted-by":"crossref","unstructured":"Perello-Nieto, M., Telmo De Menezes Filho, E.S., Kull, M., and Flach, P. (2016, January 12\u201315). Background Check: A general technique to build more reliable and versatile classifiers. Proceedings of the 2016 IEEE 16th International Conference on Data Mining (ICDM), Barcelona, Spain.","DOI":"10.1109\/ICDM.2016.0150"},{"key":"ref_24","doi-asserted-by":"crossref","unstructured":"Roady, R., Hayes, T.L., Kemker, R., Gonzales, A., and Kanan, C. (2019). Are Out-of-Distribution Detection Methods Effective on Large-Scale Datasets?. arXiv.","DOI":"10.1371\/journal.pone.0238302"},{"key":"ref_25","doi-asserted-by":"crossref","unstructured":"Nguyen, A., Yosinski, J., and Clune, J. (2015, January 7\u201312). Deep neural networks are easily fooled: High confidence predictions for unrecognizable images. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Boston, MA, USA.","DOI":"10.1109\/CVPR.2015.7298640"},{"key":"ref_26","doi-asserted-by":"crossref","first-page":"163","DOI":"10.1038\/d41586-019-03013-5","article-title":"Why deep-learning AIs are so easy to fool","volume":"574","author":"Heaven","year":"2019","journal-title":"Nature"},{"key":"ref_27","doi-asserted-by":"crossref","unstructured":"Geng, C., Sheng-jun, H., and Songcan, C. (2020). Recent advances in open set recognition: A survey. IEEE Trans. Pattern Anal. Mach. Intell.","DOI":"10.1109\/TPAMI.2020.2981604"},{"key":"ref_28","unstructured":"Gune, O., More, A., Banerjee, B., and Chaudhuri, S. (2019, January 9\u201312). Generalized Zero-shot Learning using Open Set Recognition. Proceedings of the 30th British Machine Vision Conference, Cardiff, UK."},{"key":"ref_29","doi-asserted-by":"crossref","unstructured":"Noumir, Z., Honeine, P., and Richard, C. (2012, January 1\u20136). On simple one-class classification methods. Proceedings of the 2012 IEEE International Symposium on Information Theory Proceedings, Cambridge, MA, USA.","DOI":"10.1109\/ISIT.2012.6283685"},{"key":"ref_30","doi-asserted-by":"crossref","first-page":"345","DOI":"10.1017\/S026988891300043X","article-title":"One-class classification: Taxonomy of study and review of techniques","volume":"29","author":"Khan","year":"2014","journal-title":"Knowl. Eng. Rev."},{"key":"ref_31","doi-asserted-by":"crossref","first-page":"289","DOI":"10.1109\/TNNLS.2013.2274735","article-title":"Multiclass from binary: Expanding one-versus-all, one-versus-one and ecoc-based approaches","volume":"25","author":"Rocha","year":"2013","journal-title":"IEEE Trans. Neural Netw. Learn. Syst."},{"key":"ref_32","unstructured":"J\u00fanior, P.R.M., Boult, T.E., Wainer, J., and Rocha, A. (2020). Specialized support vector machines for open-set recognition. arXiv."},{"key":"ref_33","doi-asserted-by":"crossref","unstructured":"Bendale, A., and Boult, T. (2015, January 7\u201312). Towards open world recognition. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Boston, MA, USA.","DOI":"10.1109\/CVPR.2015.7298799"},{"key":"ref_34","doi-asserted-by":"crossref","first-page":"2624","DOI":"10.1109\/TPAMI.2013.83","article-title":"Distance-based image classification: Generalizing to new classes at near-zero cost","volume":"35","author":"Mensink","year":"2013","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"ref_35","doi-asserted-by":"crossref","first-page":"359","DOI":"10.1007\/s10994-016-5610-8","article-title":"Nearest neighbors distance ratio open-set classifier","volume":"106","author":"Werneck","year":"2017","journal-title":"Mach. Learn."},{"key":"ref_36","unstructured":"Sensoy, M., Kaplan, L., and Kandemir, M. (2018). Evidential deep learning to quantify classification uncertainty. Advances in Neural Information Processing Systems, Curran Associates, Inc."},{"key":"ref_37","doi-asserted-by":"crossref","unstructured":"Shaker, M.H., and H\u00fcllermeier, E. (2020). Aleatoric and Epistemic Uncertainty with Random Forests. International Symposium on Intelligent Data Analysis, Springer.","DOI":"10.1007\/978-3-030-44584-3_35"},{"key":"ref_38","unstructured":"Depeweg, S., Jose-Miguel, H.L., Finale, D.V., and Steffen, U. (2018, January 10\u201315). Decomposition of uncertainty in Bayesian deep learning for efficient and risk-sensitive learning. Proceedings of the International Conference on Machine Learning, Stockholm, Sweden."},{"key":"ref_39","unstructured":"Shafaei, A., Schmidt, M., and Little, J.J. (2018). A less biased evaluation of out-of-distribution sample detectors. arXiv."},{"key":"ref_40","doi-asserted-by":"crossref","first-page":"4927","DOI":"10.1038\/s41467-019-12898-9","article-title":"Rapid identification of pathogenic bacteria using Raman spectroscopy and deep learning","volume":"10","author":"Ho","year":"2019","journal-title":"Nat. Commun."},{"key":"ref_41","unstructured":"Dua, D., and Graff, C. (2019). UCI Machine Learning Repository, University of California."},{"key":"ref_42","first-page":"65","article-title":"Accuracy-rejection curves (ARCs) for comparing classification methods with a reject option","volume":"8","author":"Nadeem","year":"2009","journal-title":"Mach. Learni. Syst. Biol."},{"key":"ref_43","doi-asserted-by":"crossref","unstructured":"Scott, D.W. (1992). Multivariate Density Estimation: Theory, Practice, and Visualization, John Wiley & Sons.","DOI":"10.1002\/9780470316849"},{"key":"ref_44","doi-asserted-by":"crossref","first-page":"160125","DOI":"10.1109\/ACCESS.2019.2950244","article-title":"Accuracy Rejection Normalized-Cost Curves (ARNCCs): A Novel 3-Dimensional Framework for Robust Classification","volume":"7","author":"Abbas","year":"2019","journal-title":"IEEE Access"},{"key":"ref_45","doi-asserted-by":"crossref","unstructured":"Davis, J., and Goadrich, M. (2006, January 25\u201329). The relationship between Precision-Recall and ROC curves. Proceedings of the 23rd International Conference on Machine Learning, Pittsburgh, PA, USA.","DOI":"10.1145\/1143844.1143874"},{"key":"ref_46","doi-asserted-by":"crossref","first-page":"33","DOI":"10.1007\/s10994-005-5256-4","article-title":"A response to Webb and Ting\u2019s on the application of ROC analysis to predict classification performance under varying class distributions","volume":"58","author":"Fawcett","year":"2005","journal-title":"Mach. Learn."},{"key":"ref_47","unstructured":"Hendrycks, D., and Gimpel, K. (2016). A baseline for detecting misclassified and out-of-distribution examples in neural networks. arXiv."},{"key":"ref_48","unstructured":"Manning, C., and Schutze, H. (1999). Foundations of Statistical Natural Language Processing, MIT Press."},{"key":"ref_49","doi-asserted-by":"crossref","unstructured":"Saito, T., and Rehmsmeier, M. (2015). The precision-recall plot is more informative than the ROC plot when evaluating binary classifiers on imbalanced datasets. PLoS ONE, 10.","DOI":"10.1371\/journal.pone.0118432"},{"key":"ref_50","doi-asserted-by":"crossref","first-page":"1443","DOI":"10.1162\/089976601750264965","article-title":"Estimating the support of a high-dimensional distribution","volume":"13","author":"John","year":"2001","journal-title":"Neural Comput."},{"key":"ref_51","first-page":"975","article-title":"Probability estimates for multi-class classification by pairwise coupling","volume":"5","author":"Wu","year":"2004","journal-title":"J. Mach. Learn. Res."},{"key":"ref_52","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1145\/2133360.2133363","article-title":"Isolation-based anomaly detection","volume":"6","author":"Liu","year":"2012","journal-title":"ACM Trans. Knowl. Discov. Data"}],"container-title":["Machine Learning and Knowledge Extraction"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/2504-4990\/2\/4\/28\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,10,11]],"date-time":"2025-10-11T10:27:17Z","timestamp":1760178437000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/2504-4990\/2\/4\/28"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2020,10,30]]},"references-count":52,"journal-issue":{"issue":"4","published-online":{"date-parts":[[2020,12]]}},"alternative-id":["make2040028"],"URL":"https:\/\/doi.org\/10.3390\/make2040028","relation":{},"ISSN":["2504-4990"],"issn-type":[{"type":"electronic","value":"2504-4990"}],"subject":[],"published":{"date-parts":[[2020,10,30]]}}}