Abstract: Techniques for dividing a geographical area into districts are described herein. Geospatial vector data, barrier geospatial vector data, road infrastructure data, and historical delivery demand data for a geographical area may be obtained. A plurality of clusters from a stratified sampling of data points for the delivery demand data and barrier penalties from a barrier-aware road graph are generated. A first set of polygons for the plurality of clusters may be generated using a concave hull algorithm. A second set of polygons may be generated using a barrier constrained network Voronoi algorithm that uses the barrier-aware road graph and the first set of polygons as seeds. The second set of polygons may be modified using a bounded Voronoi algorithm that uses a raster cost allocation based on barrier penalties. Coordinates for each polygon of the modified second set of polygons are determined that divide the geographical area.

Abstract: Natural language processing (NLP) methodologies and mechanisms are provided that include a named entity recognition (NER) computer model augmented to operate on an entity pattern embedding input feature in addition to other embedding input features. The mechanisms tokenize natural language content (NLC) to generate tokens and process a selected token in accordance with a predetermined entity pattern embedding technique to generate an entity pattern embedding input feature for the selected token. The entity pattern embedding input feature specifies a pattern of characters present in the selected token. The mechanisms process the NLC to generate the other embedding input features in accordance with other embedding techniques, and process, by the NER computer model, the other embedding input features and the entity pattern embedding input feature for the selected token to generate a predicted tag for the selected token. The predicted tag specifies a named entity type classification for the selected token.

Abstract: Natural language processing (NLP) methodologies and mechanisms are provided that include a named entity recognition (NER) computer model augmented to operate on an entity pattern embedding input feature in addition to other embedding input features. The mechanisms tokenize natural language content (NLC) to generate tokens and process a selected token in accordance with a predetermined entity pattern embedding technique to generate an entity pattern embedding input feature for the selected token. The entity pattern embedding input feature specifies a pattern of characters present in the selected token. The mechanisms process the NLC to generate the other embedding input features in accordance with other embedding techniques, and process, by the NER computer model, the other embedding input features and the entity pattern embedding input feature for the selected token to generate a predicted tag for the selected token. The predicted tag specifies a named entity type classification for the selected token.