Increasing NLP Parsing Efficiency with Chunking

Trátase dun resumo extendido da ponencia[Abstract] We introduce a “Chunk-and-Pass” parsing technique influenced by a psycholinguistic model, where linguistic information is processed not word-by-word but rather in larger chunks of words. We present preliminary results that show that it is feasible to compress linguistic data into chunks without significantly diminishing parsing performance and potentially increasing the speed