In the previous chapters we discussed how the meaning of words and larger grammatical constructions can be represented by machines. In 7.1 we assess, we an independent test that was built specifically for this purpose by Levesque, Davis, and Morgenstein (2012), how well these techniques stand up on semantic tasks. We develop a simple taxonomy of the 140+ tasks in this test set, and discuss their generality.
Unable to display preview. Download preview PDF.