Abstract: Systems and methods for training and utilizing predictive models that ignore missing features in accordance with embodiments of the invention are illustrated. One embodiment includes a method for generating representations of inputs with missing values. The method includes steps for, at a single layer in a multi-layer model, receiving an input includes a set of one or more values for several features and identifying a missingness pattern of the input, wherein the missingness pattern indicates whether the set of values is missing a value for each of the several features. The method further includes determining a set of one or more transformation weights based on the missingness pattern and transforming the input based on the determined transformation weights.

Abstract: A portable line barrier including a plurality of telescopically extendable poles. Further provided are a pair of line payer heads. Each line payer head is coupled to a separate pole and each head has a container with a hollow interior and a closeable opening for allowing access to the interior. A spool is disposed within the interior of the container and is rotatably coupled to the container. A plurality of loops with each loop are coupled to a separate pole which is not coupled to a line payer head. Each loop is adapted for holding a cable within. A cable which has one end coupled to and wrapped around the spool of one of the line payer heads is included. Another end extends through the loops of the poles and is coupled to and wrapped around the spool of the other line payer head such that the poles, line payer head, and loops create an extendable barrier.