Evidence-theoretical modeling of uncertainty induced by posterior probability distributions

• Autorzy: Kałuża Daniel , Janusz Andrzej , Ślęzak Dominik

Identyfikatory

DOI

10.61822/amcs-2025-0003

• Języki publikacji: EN

Abstrakty

EN

We discuss how the posterior probability distributions produced by machine learning models for analyzed objects can be transformed into evidence-theoretical mass functions that model uncertainties associated with operating those distributions. We investigate the mathematical properties of the introduced mass functions and their corresponding belief functions. We also construct some uncertainty measures based on the functions considered and compare them with several classical uncertainty measures, both theoretically and practically, in the active learning scenarios.

Słowa kluczowe

EN

theory of evidence; posterior probabilities; measures of uncertainty; active learning

• Wydawca: Oficyna Wydawnicza Uniwersytetu Zielonogórskiego
• Czasopismo: International Journal of Applied Mathematics and Computer Science
• Rocznik: 2025
• Tom: Vol. 35, no. 1
• Strony: 33--43
• Opis fizyczny: Bibliogr. 26 poz., rys., tab.

Twórcy

autor

Kałuża Daniel

• Institute of Informatics, University of Warsaw, ul. Banacha 2, 02-097 Warsaw, Poland

autor

Janusz Andrzej

• Institute of Informatics, University of Warsaw, ul. Banacha 2, 02-097 Warsaw, Poland
• School of Information Systems, Queensland University of Technology, Gardens Point Campus, Brisbane, Australia

autor

Ślęzak Dominik

• Institute of Informatics, University of Warsaw, ul. Banacha 2, 02-097 Warsaw, Poland
• QED Software, ul. Mazowiecka 11/49, 00-052 Warsaw, Poland

Bibliografia

• [1] Agrawal, A., Tripathi, S. and Vardhan, M. (2021). Active learning approach using a modified least confidence sampling strategy for named entity recognition, Progress in Artificial Intelligence 10(2): 113-128, DOI: 10.1007/s13748-021-00230-w.
• [2] Alpaydin, E. and Alimoglu, F. (1998). Pen-based recognition of handwritten digits, UCI Machine Learning Repository, https://archive.ics.uci.edu/dataset/81pen+based+recognition+of+handwritten+digits, DOI: 10.24432/C5MG6K.
• [3] Bezerra, E.D.C., Teles, A.S., Coutinho, L.R. and da Silvae Silva, F.J. (2021). Dempster-Shafer theory for modeling and treating uncertainty in IoT applications based on complex event processing, Sensors 21(5): 1863, DOI: 10.3390/s21051863.
• [4] Bohanec, M. (1997). Car evaluation, UCI Machine Learning Repository, https://archive.ics.uci.edu/dataset/19/car+evaluation, DOI: 10.24432/C5JP48.
• [5] Campagner, A., Ciucci, D. and Denoeux, T. (2022). Belief functions and rough sets: Survey and new insights, International Journal of Approximate Reasoning 143: 192-215, DOI: 10.1016/j.ijar.2022.01.011.
• [6] Cattaneo, G. (2023). Abstract approach to entropy and co-entropy in measurable and probability spaces, in M. Ganzha et al. (Eds), Proceedings of FedCSIS 2023, Annals of Computer Science and Information Systems, Vol. 35, Warsaw, pp. 63-64, DOI: 10.15439/2023F0004.
• [7] Deterding, D.H. (1990). Speaker Normalisation for Automatic Speech Recognition, PhD thesis, University of Cambridge, Cambridge.
• [8] Dubois, D. and Prade, H. (1987). Properties of measures of information in evidence and possibility theories, Fuzzy Sets and Systems 24(2): 161-182, DOI: 10.1016/0165-0114(87)90088-1.
• [9] Frey, P.W. and Slate, D.J. (1991). Letter recognition using holland-style adaptive classifiers, Machine Learning 6(2): 161-182, DOI: 10.1007/BF00114162.
• [10] Hemmer, P., Kühl, N. and Schöffer, J. (2020). DEAL: Deep evidential active learning for image classification, Proceedings International Conference on Machine Learning and Applications (ICMLA), pp. 865-870, DOI: 10.1109/ICMLA51294.2020.00141, (virtual event).
• [11] Hoarau, A., Martin, A., Dubois, J. and Gall, Y.L. (2022). Imperfect labels with belief functions for active learning, Proceedings of BELIEF 2022, Lecture Notes in Computer Science, Vol. 13506, Springer, Cham, pp. 44-53, DOI: 10.1007/978-3-031-17801-6_5.
• [12] Hunter, J.D. (2007). Matplotlib: A 2D graphics environment, Computing in Science & Engineering 9(3): 90-95, DOI: 10.1109/MCSE.2007.55.
• [13] Ikeda, Y. (2024). yuzie007/mpltern: 1.0.4., https://zenodo.org/records/11068993, DOI: 10.5281/zenodo.11068993.
• [14] Janusz, A., Zalewska, A., Wawrowski, Ł., Biczyk, P., Ludziejewski, J., Sikora, M. and Ślęzak, D. (2023). BrightBox - A rough set based technology for diagnosing mistakes of machine learning models, Applied Soft Computing 141: 110285, DOI: 10.1016/j.asoc.2023.110285.
• [15] Kałuża, D., Janusz, A. and Ślęzak, D. (2023a). On several new Dempster-Shafer-inspired uncertainty measures applicable for active learning, in A. Campagner et al. (Eds), Proceedings of IJCRS 2023, Lecture Notes in Computer Science, Vol. 14481, Springer, Cham, pp. 479-494, DOI: 10.1007/978-3-031-50959-9_33.
• [16] Kałuża, D., Janusz, A. and Ślęzak, D. (2023b). Robust assignment of labels for active learning with sparse and noisy annotations, Proceedings of ECAI 2023, in K. Gal et al. (Eds), Frontiers in Artificial Intelligence and Applications, IOS Press, Amsterdam, pp. 1207-1214, DOI: 10.3233/FAIA230397.
• [17] Nguyen, V., Shaker, M.H. and Hüllermeier, E. (2022). How to measure uncertainty in uncertainty sampling for active learning, Machine Learning 111(1): 89-122, DOI: 10.1007/s10994-021-06003-9.
• [18] Pedregosa, F., Varoquaux, G., Gramfort, A., Michel, V., Thirion, B., Grisel, O., Blondel, M., Prettenhofer, P., Weiss, R., Dubourg, V., VanderPlas, J., Passos, A., Cournapeau, D., Brucher, M., Perrot, M. and Duchesnay, E. (2011). Scikit-learn: Machine learning in Python, Journal of Machine Learning Research 12(85): 2825-2830.
• [19] Pięta, P. and Szmuc, T. (2021). Applications of rough sets in big data analysis: An overview, International Journal of Applied Mathematics and Computer Science 31(4): 659-683, DOI: 10.34768/amcs-2021-0046.
• [20] Scheffer, T., Decomain, C. and Wróbel, S. (2001). Active hidden Markov models for information extraction, in F. Hoffmann et al. (Eds), Proceedings of IDA 2001, Lecture Notes in Computer Science, Vol. 2189, Springer, Berlin/Heidelberg, pp. 309-318, DOI: 10.1007/3-540-44816-0_31.
• [21] Settles, B. (2012). Active Learning, Morgan & Claypool, San Rafael, DOI: 10.2200/S00429ED1V01Y201207AIM018.
• [22] Ślęzak, D. (2002). Approximate Decision Reducts, PhD thesis, University of Warsaw, Warsaw, (in Polish).
• [23] Smets, P. (2005). Decision making in the TBM: The necessity of the pignistic transformation, International Journal of Approximate Reasoning 38(2): 133-147, DOI: 10.1016/j.ijar.2004.05.003.
• [24] Vandoni, J., Aldea, E. and Le Hégarat-Mascle, S. (2019). Evidential query-by-committee active learning for pedestrian detection in high-density crowds, International Journal of Approximate Reasoning 104: 166-184, DOI: 10.1016/j.ijar.2018.11.007.
• [25] Yager, R.R. and Liu, L. (2008). Classic Works of the Dempster-Shafer Theory of Belief Functions, Springer, Berlin/Heidelberg.
• [26] Zhang, G. (2021). Four uncertain sampling methods are superior to random sampling method in classification, Proceedings of ICAIE 2021, Dali, China, pp. 209-212.

Uwagi

Opracowanie rekordu ze środków MNiSW, umowa nr POPUL/SP/0154/2024/02 w ramach programu "Społeczna odpowiedzialność nauki II" - moduł: Popularyzacja nauki (2025).