open access

https://hdl.handle.net/2115/93033

Extending Binary Neural Networks to Bayesian Neural Networks with Probabilistic Interpretation of Binary Weights

英語

uncertainty estimation

2024-08-01

IEICE - Institute of the Electronics, Information and Communication Engineers

IEICE transactions on information and systems

E107D(8)

949-957

Neural networks (NNs) fail to perform well or make excessive predictions when predicting out-of-distribution or unseen datasets. In contrast, Bayesian neural networks (BNNs) can quantify the uncertainty of their inference to solve this problem. Nevertheless, BNNs have not been widely adopted owing to their increased memory and computational cost. In this study, we propose a novel approach to extend binary neural networks by introducing a probabilistic interpretation of binary weights, effectively converting them into BNNs. The proposed approach can reduce the number of weights by half compared to the conventional method. A comprehensive comparative analysis with established methods like Monte Carlo dropout and Bayes by backprop was performed to assess the performance and capabilities of our proposed technique in terms of accuracy and capturing uncertainty. Through this analysis, we aim to provide insights into the advantages of this Bayesian extension.

学術雑誌論文

Version of Record

https://search.ieice.org/

Copyright ©2024 The Institute of Electronics, Information and Communication Engineers

0916-8532