Adaptive sampling using variational autoencoder and reinforcement learning
By: Adil Rasheed, Mikael Aleksander Jansen Shahly, Muhammad Faisal Aftab
Potential Business Impact:
Finds hidden patterns with fewer guesses.
Compressed sensing enables sparse sampling but relies on generic bases and random measurements, limiting efficiency and reconstruction quality. Optimal sensor placement uses historcal data to design tailored sampling patterns, yet its fixed, linear bases cannot adapt to nonlinear or sample-specific variations. Generative model-based compressed sensing improves reconstruction using deep generative priors but still employs suboptimal random sampling. We propose an adaptive sparse sensing framework that couples a variational autoencoder prior with reinforcement learning to select measurements sequentially. Experiments show that this approach outperforms CS, OSP, and Generative model-based reconstruction from sparse measurements.
Similar Papers
A Novel Monte-Carlo Compressed Sensing and Dictionary Learning Method for the Efficient Path Planning of Remote Sensing Robots
Robotics
Robot drones map pollution using less travel.
Compressed Sensing: Mathematical Foundations, Implementation, and Advanced Optimization Techniques
Machine Learning (CS)
Gets more information from fewer signals.
Probabilistic and nonlinear compressive sensing
Machine Learning (CS)
Finds best data patterns faster than other methods.