Kabir, HM Dipu, Mondal, Subrota Kumar, Khanam, Sadia, Khosravi, Abbas, Rahman, Shafin, Chalak Qazani, Mohammad Reza, Alizadehsani, Roohallah, Asadi, Houshyar, Mohamed, Shady, Nahavandi, Saeid, and Acharya, U. Rajendra (2023) Uncertainty aware neural network from similarity and sensitivity. Applied Soft Computing, 149 (Part A). 111027.

PDF (Published Version) - Published Version Restricted to Repository staff only

Abstract

Recent uncertainty quantification approaches lack transparency. Algorithms often perform poorly in an input domain and the reason for poor performance remains unknown. Therefore, we present a neural network training method that considers similar samples with sensitivity awareness in this paper. In the proposed NN training method for UQ, first, we train a shallow NN for the point prediction. Then, we train another NN to predict absolute differences between targets and predictions. In the next step, we select each sample in the training set one by one and compute both prediction and error sensitivities. Then we select similar samples with sensitivity consideration and save indexes of similar samples. The ranges of an input parameter become narrower when the output is highly sensitive to that parameter. After that, we construct initial uncertainty bounds (UB) by considering the distribution of sensitivity-aware similar samples. Prediction intervals (PIs) from initial uncertainty bounds are larger and cover more samples than required. Therefore, we train bound correction NN. As following all the steps for finding UB for each sample requires a lot of computation and memory access, we train a UB computation NN. The UB computation NN takes an input sample and provides an uncertainty bound. The UB computation NN is the final product of the proposed approach. We have achieved superior performance in most situations through the proposed transparent method.

Item ID:	86721
Item Type:	Article (Research - C1)
ISSN:	1872-9681
Copyright Information:	© 2023 Elsevier B.V. All rights reserved.
Funders:	Australian Research Council (ARC)
Projects and Grants:	ARC Discovery Project funding scheme project DP190102181
Date Deposited:	10 Sep 2025 01:24
FoR Codes:	46 INFORMATION AND COMPUTING SCIENCES > 4611 Machine learning > 461104 Neural networks @ 60% 46 INFORMATION AND COMPUTING SCIENCES > 4601 Applied computing > 460199 Applied computing not elsewhere classified @ 40%
SEO Codes:	28 EXPANDING KNOWLEDGE > 2801 Expanding knowledge > 280115 Expanding knowledge in the information and computing sciences @ 100%
	More Statistics