Redundant Information Neural Estimation

We introduce the Redundant Information Neural Estimator (RINE), a method that allows efficient estimation for component of information about target variable is common to set sources, known as “redundant information”. show existing definitions redundant can be recast in terms an optimization over family functions. In contrast previous decompositions, which only evaluated discrete variables small alphabets, we optimizing functions enables approximation high-dimensional and continuous predictors. demonstrate this on image classification motor-neuroscience tasks.