The Two-Way Cooperative Information Bottleneck

MATÍAS ALEJANDRO VERA; LEONARDO REY VEGA; PABLO PIANTANIDA

Simposio; IEEE International Symposium on Information Theory (ISIT); 2015

The two-way Information Bottleneck problem,where two nodes exchange information iteratively about twoarbitrarily dependent memoryless sources, is considered. Basedon the observations and the information exchange, each node isrequired to extract ?relevant information?, measured in termsof the normalized mutual information, from two arbitrarilydependent hidden sources. The optimal trade-off between ratesof relevance and complexity, and the number of exchange rounds,is obtained through a single-letter characterization. We furtherextend the results to the Gaussian case. Applications of our setuparise in the development of collaborative clustering algorithms.