NGO-GM: Natural Gradient Optimization for Graphical Models
This provides a novel optimization method for graphical models, offering an alternative to EM with potential applications in domains like finance, though it appears incremental as it builds on existing natural gradient concepts.
The paper tackles parameter estimation in graphical models by reformulating it as an information geometric optimization problem and introducing a natural gradient descent strategy with meta parameters. The result shows that this approach outperforms traditional practitioner methods in financial trend detection, being less prone to overfitting.
This paper deals with estimating model parameters in graphical models. We reformulate it as an information geometric optimization problem and introduce a natural gradient descent strategy that incorporates additional meta parameters. We show that our approach is a strong alternative to the celebrated EM approach for learning in graphical models. Actually, our natural gradient based strategy leads to learning optimal parameters for the final objective function without artificially trying to fit a distribution that may not correspond to the real one. We support our theoretical findings with the question of trend detection in financial markets and show that the learned model performs better than traditional practitioner methods and is less prone to overfitting.