Loading...
Please wait, while we are loading the content...
Similar Documents
Variational Inference for Robust Sequential Learning of Multilayered Perceptron Neural Network
| Content Provider | Semantic Scholar |
|---|---|
| Author | Vukovi, Najdan Miti, Marko Miljkovi, Zoran |
| Copyright Year | 2015 |
| Abstract | We derive a new sequential learning algorithm for Multilayered Perceptron (MLP) neural network robust to outliers. Presence of outliers in data results in failure of the model especially if data processing is performed on-line or in real time. Extended Kalman filter robust to outliers (EKF-OR) is probabilistic generative model in which measurement noise covariance is modeled as stochastic process over the set of symmetric positive-definite matrices in which prior is given as inverse Wishart distribution. Derivation of expressions comes straight form first principles, within Bayesian framework. Analytical intractability of Bayes’ update step is solved using Variational Inference (VI). Experimental results obtained using real world stochastic data show that MLP network trained with proposed algorithm achieves low error and average improvement rate of 7% when compared directly to conventional EKF learning algorithm. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://www.mas.bg.ac.rs/_media/istrazivanje/fme/vol43/2/6_nvukovic.pdf |
| Language | English |
| Access Restriction | Open |
| Subject Keyword | Algorithm Approximation Artificial neural network Data logger Data point Engineering Estimated Extended Kalman filter Formal Methods Europe Foundations Generalization (Psychology) Generative model Inference Intuition Memory-level parallelism Online and offline Outlier Perceptron Quad Flat No-leads package Real-time computing Sequential algorithm Stochastic process Theoretical definition Time series Variational inequality Variational principle |
| Content Type | Text |
| Resource Type | Article |
