Sample-optimal and efficient learning of tree Ising models

We show that n-variable tree-structured Ising models can be learned computationally-efficiently to within total variation distance from an optimal O(n ln n/ 2 ) samples, where O(•) hides an absolute constant which, importantly, does not depend on the model being learned-neither its tree nor the magnitude of its edge strengths, on which we place no assumptions. Our guarantees hold, in fact, for the celebrated Chow-Liu algorithm [5], using the plug-in estimator for estimating mutual information. While this (or any other) algorithm may fail to identify the structure of the underlying model correctly from a finite sample, we show that it will still learn a tree-structured model that is -close to the true one in total variation distance, a guarantee called "proper learning."

Our guarantees do not follow from known results for the Chow-Liu algorithm [6] and the ensuing literature on learning graphical models, including the very recent renaissance of algorithms on this learning challenge (see e.g. [2,21,13,10,24,22]), which only yield asymptotic consistency results, or sampleinefficient and/or time-inefficient algorithms, unless further assumptions are placed on the graphical model, such as bounds on the "strengths" of the model's edges/hyperedges. While we establish guarantees for a widely known and simple algorithm, the analysis that this algorithm succeeds and is sample-optimal is quite complex, requiring a hierarchical classification of the edges into layers with different reconstruction guarantees, depending on their strength, combined with delicate uses of the subadditivity of the squared Hellinger distance over graphical models to control the error accumulation.