Abstract: Embodiments of the present invention relate to performing entity recognition on a stream while providing ongoing training or supplementation of an entity dictionary. In one embodiment, a method of and computer program product for stream based named entity recognition is provided. A first portion of a textual input is received. A plurality of patterns is applied to the first portion to determine that a predetermined type is present in the first portion. Approval is requested of the presence of the predetermined type. An indication of approval or disapproval of the predetermined type is received. A dictionary is supplemented according to the indication. A second portion of the textual input is received. The plurality of patterns is applied to the second portion.

Abstract: A method for updating automated annotations for a paper-based document is provided. The method may include receiving an image of the paper-based document, the image of the paper-based document including a plurality of highlighted annotations having at least one edited annotation. The method may also include extracting the at least one edited annotation from the received image of the paper-based document. The method may further include adding the extracted at least one edited annotation of the paper-based document to a data retention system.

Abstract: Embodiments of the present invention relate to interactive optimization of messages published to digital media based on past performance of similar messages. In one embodiment, an input token is received. At least one candidate substitute token is retrieved from a dictionary. The dictionary comprises a mapping from the input token to the at least one candidate substitute token. A score associated with the at least one candidate substitute token is determined. A score associated with the input token is determined. The score associated with the input token, the at least one candidate substitute token, and the score associated with the at least one candidate substitute token are outputted.

Abstract: Embodiments of the present invention relate to scoring of messages published to digital media based on past performance of similar messages. In one embodiment, an input token is received. A plurality of messages is selected from a corpus of messages. Each of the plurality of messages has a publication time and contents. The contents of each of the plurality of messages include the input token. A plurality of root messages is determined from the plurality of messages. Each of the plurality of root messages relates to at least one related message. The at least one related message is one of the plurality of messages. Each of the plurality of root messages is the earliest message of the corpus of messages related to its at least one related message. A score is determined for the input token based on the plurality of root messages.

Abstract: A method and apparatus for selectively removing fine sediments from a water body bottom using a dredging head with one or more intake openings situated to accelerate flows through the interior of the head and across the water body bottom causing fine particles to become suspended in a slurry outflowing from the head for disposal, while coarser, heavier, bottom gravels remain in place. The stream cleaner dredging head includes a main body having at least one intake, at least one plate attachment feature, a discharge point, and an at least partly open bottom; a flow-adjusting plate that is adjustably connected to the at least one plate attachment feature and configured to adjust a fluid flow through the main body from the at least one intake, and a hose attachment connected to the discharge point of the main body.

Abstract: A golf disc assembly includes a disc configured for flying upon being cast into the air and a sonic finding unit which is attached to the disc and which emits a sonic locating signal after the sonic finding unit conducts a delay of a predetermined amount of time.