Combining Vector Space Embeddings with Symbolic Logical Inference over Open-Domain Text.
Matt Gardner, Partha Talukdar, Tom M. Mitchell · 2015
We have recently shown how to combine random walk inference over knowledge bases with vector space rep-resentations of surface forms, improving performance on knowledge base inference. In this paper, we formal-ize the connection of our prior work to logical inference rules, giving some general observations about meth-ods for incorporating vector space representations into symbolic logic systems. Additionally, we present some promising preliminary work that extends these tech-niques to learning open-domain relations for the pur-pose of answering multiple choice questions, achieving 67 % accuracy on a small test set.