Conference proceeding
Approximate Bayesian Neural Network Trained with Ensemble Kalman Filter
2019 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS (IJCNN), Vol.2019-, pp.1-8
IEEE International Joint Conference on Neural Networks (IJCNN)
01/01/2019
DOI: 10.1109/IJCNN.2019.8851742
Abstract
Neural networks have achieved significant success in many areas. Nevertheless, conventional neural networks lack uncertainty information, which plays an important role especially in critical-safety applications such as self-driving cars. When uncertainty is characterized using probability the main modeling approach is the construction of Bayesian neural networks. Obtaining the posterior distribution for these models is computationally intensive and analytical solutions are intractable. In this work, we propose a novel algorithm to infer the weights for Bayesian neural networks based on the ensemble Kalman filter. To evaluate the performance of the algorithm, we use ten regression datasets from University of California at Irvine machine learning repository, and a natural language dataset. The results suggest that EnKF can be used as a gradient-free alternative to training deep neural networks to capture prediction uncertainty.
Details
- Title: Subtitle
- Approximate Bayesian Neural Network Trained with Ensemble Kalman Filter
- Creators
- Chao Chen - University of South CarolinaXiao Lin - University of South CarolinaYuan Huang - University of South CarolinaGabriel Terejanu - University of North Carolina at Charlotte
- Resource Type
- Conference proceeding
- Publication Details
- 2019 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS (IJCNN), Vol.2019-, pp.1-8
- Publisher
- IEEE
- Series
- IEEE International Joint Conference on Neural Networks (IJCNN)
- DOI
- 10.1109/IJCNN.2019.8851742
- ISSN
- 2161-4393
- eISSN
- 2161-4407
- Number of pages
- 8
- Grant note
- 2017-67017-26167 / National Institute of Food and Agriculture (NIFA)/USDA
- Language
- English
- Date published
- 01/01/2019
- Academic Unit
- Biostatistics
- Record Identifier
- 9984364418002771
Metrics
8 Record Views