Natural language processing is as far fromas bricks are from a house…. As far as I know, there is no benchmark for NLU, other than perhaps the much maligned .
In fact, DARPA has recently launched a new program for new benchmarks, called. Previously, DARPA had announced the “Machine Reading Program”, subsequently renamed , apparently in effect through July 2017.
I believe that onlyis a true test of Natural language understanding. Therefore, the current batch of NLG companies represent “state of the art” in making sense of raw data, and are listed below. Further, I maintain that “stories, articles or paragraphs” must first be broken down into (raw) data, which is more or less what NLP does….
(Visited 15 times, 1 visits today)