Statistical Relational Learning to Recognise Textual Entailment

Miguel Angel Rios Gaona

Abstract:

We propose a novel approach to recognise
textual entailment (RTE) following a twostage architecture – alignment and decision – where both stages are based on semantic representations. In the alignment
stage the entailment candidate pairs are
represented and aligned using predicateargument structures. In the decision stage,
a Markov Logic Network (MLN) is learnt
using rich relational information from the
alignment stage to predict an entailment
decision. We evaluate this approach using the RTE Challenge datasets. It shows
comparable results against the average
performance across participating systems,
and very promising results for a subset of
the datasets for which a semantic alignment can be found, evidencing the potential of MLNs for RTE