The information bottleneck (IB) approach to clustering takes a joint distribution P(X,Y) and maps the data X to cluster labels T , which retain maximal information about Y (Tishby, Pereira, & Bialek, 1999 ). This objective results in an algorithm that clusters data points based on the similarity of their conditional distributions P(Y∣X). This is in contrast to classic geometric clustering algorithms such as k -means and gaussian mixture models (GMMs), which take a set of observed data points {xi}i=1:N and cluster them based on their geometric (typically Euclidean) distance from one another. ⋯ We apply this approach to a variety of simple clustering problems, showing that DIB with our model selection procedure recovers the generative cluster labels. We also show that, in particular limits of our model parameters, clustering with DIB and IB is equivalent to k -means and EM fitting of a GMM with hard and soft assignments, respectively. Thus, clustering with (D)IB generalizes and provides an information-theoretic perspective on these classic algorithms.
Department of Physics, Princeton University, Princeton, NJ 08544, U.S.A. danieljstrouse@gmail.com.
Neural Comput. 2019 Mar 1; 31 (3): 596-612.
AbstractThe information bottleneck (IB) approach to clustering takes a joint distribution http://www.w3.org/1998/Math/MathML">P(X,Y) and maps the data http://www.w3.org/1998/Math/MathML">X to cluster labels http://www.w3.org/1998/Math/MathML">T , which retain maximal information about http://www.w3.org/1998/Math/MathML">Y (Tishby, Pereira, & Bialek, 1999 ). This objective results in an algorithm that clusters data points based on the similarity of their conditional distributions http://www.w3.org/1998/Math/MathML">P(Y∣X) . This is in contrast to classic geometric clustering algorithms such as http://www.w3.org/1998/Math/MathML">k -means and gaussian mixture models (GMMs), which take a set of observed data points http://www.w3.org/1998/Math/MathML">{xi}i=1:N and cluster them based on their geometric (typically Euclidean) distance from one another. Here, we show how to use the deterministic information bottleneck (DIB) (Strouse & Schwab, 2017 ), a variant of IB, to perform geometric clustering by choosing cluster labels that preserve information about data point location on a smoothed data set. We also introduce a novel intuitive method to choose the number of clusters via kinks in the information curve. We apply this approach to a variety of simple clustering problems, showing that DIB with our model selection procedure recovers the generative cluster labels. We also show that, in particular limits of our model parameters, clustering with DIB and IB is equivalent to http://www.w3.org/1998/Math/MathML">k -means and EM fitting of a GMM with hard and soft assignments, respectively. Thus, clustering with (D)IB generalizes and provides an information-theoretic perspective on these classic algorithms.