Gaussian Attention Model and Its Application to Knowledgebase Embedding and Question Answering. (arXiv:1611.02266v1 [cs.CL])

We propose the Gaussian attention model for content-based neural memory access. With the proposed attention model, a neural network has the additional degree of freedom to control the focus of its attention from a laser sharp attention to blurred attention. It is applicable whenever we can assume that the distance in the latent space reflects some notion of semantics. We use the proposed attention model as a scoring function for the embedding of a knowledgebase into a continuous vector space and then train a model that performs question answering about the entities in the knowledgebase. The proposed attention model can handle both the propagation of uncertainty when following a series of relations and also the conjunction of thoughts in a natural way. On a dataset of soccer players who participated in the FIFA World Cup 2014, we demonstrate that our model can handle both path queries and conjunctive queries well.

from cs.AI updates on arXiv.org http://ift.tt/2fiAy5z
via IFTTT