Learn More
We present a simple EM-based grammar induction algorithm for Combinatory Categorial Grammar (CCG) that achieves state-of-the-art performance by relying on a minimal number of very general linguistic principles. Unlike previous work on unsupervised parsing with CCGs, our approach has no prior language-specific knowledge, and discovers all categories(More)
Work in grammar induction should help shed light on the amount of syntactic structure that is discoverable from raw word or tag sequences. But since most current grammar induction algorithms produce unlabeled dependencies, it is difficult to analyze what types of constructions these algorithms can or cannot capture, and, therefore, to identify where(More)
Nearly all work in unsupervised grammar induction aims to induce unlabeled dependency trees from gold part-of-speech-tagged text. These clean linguistic classes provide a very important, though unreal-istic, inductive bias. Conversely, induced clusters are very noisy. We show here, for the first time, that very limited human supervision (three frequent(More)
In this research we analyze the task of evolving a neural network to understand simple English commands. By understand we mean that the final agent will perform tasks and interact with objects in its world as instructed by the experimenter. The lexicon and grammar are kept small in this experiment. This type of work where semantics are based on an agent's(More)