.png){kind=icon}
.png){kind=icon}
Return
Neural network implementation of inference on binary Markov random fields with probability coding
.png){kind=icon}
PRE
AI
DOI：10.1016/j.amc.2016.12.025.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Markov random fields (MRF) underpin the solution to many problems in computational neuroscience. However, how the inference for MRF could be implemented with neural network is still an important open question. In this paper, we build the relationship between inference equation of MRF and the dynamic equation of the Hopfield network with probability coding. We prove that the membrane potential in the Hopfield network varying with respect to time can implement marginal probabilities inference on binary MRF. Theoretical analysis and experimental results show that our neural network can get comparable results as that of loopy belief propagation (LBP). (C) 2016 Elsevier Inc. All rights reserved.
Keywords:
Markov random fields
Approximate inference
Neural network implementation
Hopfield network
Probability coding
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
3.4
Papers:
2.3W
Citations:
3.3W