Hyper-Parameter Auto-Tuning for Sparse Bayesian Learning
Preprint
Abstract
Choosing the values of hyper-parameters in sparse Bayesian learning (SBL) can significantly impact performance. However, the hyper-parameters are normally tuned manually, which is often a difficult task. Most recently, effective automatic hyper-parameter tuning was achieved by using an empirical auto-tuner. In this work, we address the issue of hyper-parameter auto-tuning using neural network (NN)-based learning. Inspired by the empirical auto-tuner, we design and learn a NN-based auto-tuner, and show that considerable improvement in convergence rate and recovery performance can be achieved.Keywords
All Related Versions
This publication has 8 references indexed in Scilit:
- Unitary Approximate Message Passing for Sparse Bayesian LearningIEEE Transactions on Signal Processing, 2021
- Algorithm Unrolling: Interpretable, Efficient Deep Learning for Signal and Image ProcessingIEEE Signal Processing Magazine, 2021
- Sparse Bayesian Learning for Robust PCA: Algorithms and AnalysesIEEE Transactions on Signal Processing, 2019
- Sparse Bayesian Learning Based on Approximate Message Passing with Unitary Transformation2019
- Efficient and Robust Recovery of Sparse Signal and Image Using Generalized Nonconvex RegularizationIEEE Transactions on Computational Imaging, 2017
- Structured Compressed Sensing: From Theory to ApplicationsIEEE Transactions on Signal Processing, 2011
- Near-Optimal Signal Recovery From Random Projections: Universal Encoding Strategies?IEEE Transactions on Information Theory, 2006
- An Introduction to Variational Methods for Graphical ModelsMachine Learning, 1999