In June, members of the HD-NLP Group at ICL are participating in the
17th International Conference on Computational Semantics (IWCS 2023) in Nancy, with two publications and an invited talk.
SMARAGD: Learning SMatch for Accurate and Rapid Approximate Graph Distance,
by Juri Opitz, Philipp Meier and Anette Frank, to appear in IWCS 2023We show how graph distance, a hard combinatorial problem, can be accurately approximated with neural networks and data augmentation.
AMR4NLI: Interpretable and robust NLI measures from semantic graphs,
by Juri Opitz, Shira Wein, Julius Steen, Anette Frank and Nathan Schneider, to appear in IWCS 2023This paper shows that linguistic representations such as AMR can help increase the interpretability, robustness, and accuracy of NLI predictions.
Congratulations to all authors!