An implementation of deterministic annealing for clustering. My goal is to understand phase transitions in clustering, while annealing. My broader motivation is to understand this as an application of the Information Bottleneck principle and renormalization group flow, and the implications for deep learning.
- Statistical mechanics and phase transitions in clustering by Kenneth Rose, Eitan Gurewitz, and Geoffrey C. Fox
- Deterministic Annealing for Clustering, Compression, Classification, Regression, and Related Optimization Problem by Kenneth Rose