Abstract:
A method of processing limited natural language data to automatically develop an optimal feature set (510), bypassing the standard Wizard of OZ (WOZ) approach is provided. The method provides for building natural language understanding models (512) or for processing existing data from other domains, such as the Internet, for domain-specific adaptation through the use of an optimal feature set. Consequently, when the optimal feature set is passed on to any engine, the optimal feature set produces robust models that can be used for natural language call routing.
Abstract:
An embodiment of the invention is associated with a system (100) disposed to post software related tasks to a crowdsourcing marketplace (108) for execution. Prior to posting a specified task, it is determined whether the specified task comprises a complex task (306). Responsive to determining that the specified task comprises a complex task, the specified task is evaluated with respect to a first criterion, to determine whether the specified task should be decomposed into two or more atomic tasks (308). Responsive to determining that the specified task does not comprise a complex task, the specified task is evaluated with respect to a second criterion, to determine whether the specified task should be combined with other tasks into a bundled task (310).