Analyzing Training Using Phase Transitions in Entropy---Part I: General Theory
arXiv:2012.00970
Abstract
We analyze phase transitions in the conditional entropy of a sequence caused by a change in the conditional variables. Such transitions happen, for example, when training to learn the parameters of a system, since the transition from the training phase to the data phase causes a discontinuous jump in the conditional entropy of the measured system response. For large-scale systems, we present a method of computing a bound on the mutual information obtained with one-shot training, and show that this bound can be calculated using the difference between two derivatives of a conditional entropy. The system model does not require Gaussianity or linearity in the parameters, and does not require worst-case noise approximations or explicit estimation of any unknown parameters. The model applies to a broad range of algorithms and methods in communication, signal processing, and machine learning that employ training as part of their operation.
References in corpus (6)
- Channel Estimation and Performance Analysis of One-Bit Massive MIMO Systems
- Spin-Glass Theory for Pedestrians
- Optimal Errors and Phase Transitions in High-Dimensional Generalized Linear Models
- Statistical Physics of Signal Estimation in Gaussian Noise: Theory and Examples of Phase Transitions
- On an Achievable Rate of Large Rayleigh Block-Fading MIMO Channels with No CSI
- Phase Transitions for the Information Bottleneck in Representation Learning