作者: Georgiana Dinu , Marco Baroni
DOI:
关键词:
摘要: We introduce DISSECT, a toolkit to build and explore computational models of word, phrase sentence meaning based on the principles distributional semantics. The focuses in particular compositional meaning, implements number composition methods that have been proposed literature. Furthermore, DISSECT can be useful researchers practitioners who need word (without composition) as well, it supports various construct semantic spaces, assessing similarity even evaluating against benchmarks, are independent infrastructure.