Abstract: Methods, systems, and devices to support intent discovery using natural language interface are described. A computing system associated with an organization may perform intent discovery using a neuro-symbolic method to generate a set of intents for a set of conversation data sets. For example, the computing system may identify a set of hypotheses for a set of conversation data sets using domain-specific information. The computing system may use a natural language inference engine to generate, for each of a set of data pairs which each include a respective conversation data set and a respective hypothesis, a probability of whether the hypothesis of the data pair is positively entailed by the conversation data set of the data pair. For each probability that satisfies a threshold probability, the computing system may add the associated hypothesis to a set of generated intents.

Abstract: This invention is a control system for controlling an electrostatic printer/plotter electrostatically generating an image on a recording media being driven by a transport motor over a plurality of image-forming electrodes by applying image writing pulses to the electrodes. There is a media sensor for sensing which one of a plurality of pre-determined media types is currently being used in the printer/plotter and a sensor for measuring a parameter related to conductivity of the media. There is also a microprocessor for determinnig from the conductivity and type of media currently being used in the printer/plotter an optimum pulse width for the image writing pulses which maintains the density of an image produced on the media by the electrodes at a maximum density and for generating images employing image writing pulses transmitted to the electrodes which are of the optimum pulse width. Pulse repetition rate and media speed are also regulated in an optimum manner.