| Autor: |
Andrews, Emad A. M., Bonner, Anthony J. |
| Zdroj: |
Neural Information Processing (9783642106767); 2009, p100-109, 10p |
| Abstrakt: |
Belief revision is the problem of finding the most plausible explanation for an observed set of evidences. This has many applications in various scientific domains like natural language understanding, medical diagnosis and computational biology. Bayesian Networks (BN) is an important probabilistic graphical formalism used widely for belief revision tasks. In BN, belief revision can be achieved by setting the values of all random variables such that their joint probability is maximized. This assignment is called the maximum a posteriori (MAP) assignment. Finding MAP is an NP-Hard problem. In this paper, we are proposing finding the MAP assignment in BN using High Order Recurrent Neural Networks through an intermediate representation of Cost-Based Abduction. This method will eliminate the need to explicitly construct the energy function in two steps, objective and constraints, which will decrease the number of free parameters to set. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
