Abstract: The present disclosure provides a generative modeling framework for generating highly realistic and privacy preserving synthetic records for heterogenous time-series data, such as electronic health record data, financial data, etc. The generative modeling framework is based on a two-stage model that includes sequential encoder-decoder networks and generative adversarial networks (GANs).

Abstract: An alfalfa separator is provided for separating the coarser stem and fiber portion of the alfalfa plant from the finer leaf portion. The separator includes a housing having a chamber therein. The bottom part of the chamber is defined by a screen which is sized to pass the finer portions of the alfalfa. A drum having a plurality of hammers located about the periphery of the drum is mounted for rotation within the chamber. The drum serves, by means of the hammers and the centrifugal forces generated during rotation, to grind up the alfalfa so that the finer portions pass through the screen. A passageway, located above the screen and including an inlet aperture in communication with the chamber, is disposed so that the coarser portions can pass out from the chamber under the influence of the forces referred to above. A gate controls opening and closing of the passage, and suction is used in drawing off and removing the coarser portions.