期刊
ENTROPY
卷 24, 期 11, 页码 -出版社
MDPI
DOI: 10.3390/e24111675
关键词
Poisson process; renewal process; hidden semi-Markov process; hidden Markov chain; epsilon-machine; Shannon entropy rate; optimal predictor; minimal predictor
资金
- Templeton World Charity Foundation Diverse Intelligences grant [TWCF0570]
- Foundational Questions Institute and Fetzer Franklin Fund [FQXI-RFP-CPW-2007]
- U.S. Army Research Laboratory
- U.S. Army Research Office [W911NF-21-1-0048, W911NF-18-1-0028]
- U.S. Department of Energy [DE-SC0017324]
- Moore Foundation
- U.S. Department of Energy (DOE) [DE-SC0017324] Funding Source: U.S. Department of Energy (DOE)
This paper proposes new methods for inferring, predicting, and estimating continuous-time discrete-event processes. The methods are based on an extension of Bayesian structural inference and utilize the universal approximation power of neural networks. Experimental results on complex synthetic data demonstrate that these methods are competitive with the state-of-the-art for prediction and entropy-rate estimation.
Inferring models, predicting the future, and estimating the entropy rate of discrete-time, discrete-event processes is well-worn ground. However, a much broader class of discrete-event processes operates in continuous-time. Here, we provide new methods for inferring, predicting, and estimating them. The methods rely on an extension of Bayesian structural inference that takes advantage of neural network's universal approximation power. Based on experiments with complex synthetic data, the methods are competitive with the state-of-the-art for prediction and entropy-rate estimation.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据