
Using Artificial Neural Network Condensation to Facilitate Adaptation of Machine Learning in Medical Settings by Reducing Computational Burden: Model Design and Evaluation Study
Author(s) -
Dianbo Liu,
Ming Zheng,
Nestor A. Sepulveda
Publication year - 2021
Publication title -
jmir formative research
Language(s) - English
Resource type - Journals
ISSN - 2561-326X
DOI - 10.2196/20767
Subject(s) - computer science , machine learning , artificial intelligence , recurrent neural network , artificial neural network , pruning , reduction (mathematics) , computational complexity theory , algorithm , geometry , mathematics , agronomy , biology
Background Machine learning applications in the health care domain can have a great impact on people’s lives. At the same time, medical data is usually big, requiring a significant number of computational resources. Although this might not be a problem for the wide adoption of machine learning tools in high-income countries, the availability of computational resources can be limited in low-income countries and on mobile devices. This can limit many people from benefiting from the advancement in machine learning applications in the field of health care. Objective In this study, we explore three methods to increase the computational efficiency and reduce model sizes of either recurrent neural networks (RNNs) or feedforward deep neural networks (DNNs) without compromising their accuracy. Methods We used inpatient mortality prediction as our case analysis upon review of an intensive care unit dataset. We reduced the size of RNN and DNN by applying pruning of “unused” neurons. Additionally, we modified the RNN structure by adding a hidden layer to the RNN cell but reducing the total number of recurrent layers to accomplish a reduction of the total parameters used in the network. Finally, we implemented quantization on DNN by forcing the weights to be 8 bits instead of 32 bits. Results We found that all methods increased implementation efficiency, including training speed, memory size, and inference speed, without reducing the accuracy of mortality prediction. Conclusions Our findings suggest that neural network condensation allows for the implementation of sophisticated neural network algorithms on devices with lower computational resources.