Context-aware Entity Morph Decoding

  • Boliang Zhang ,
  • Hongzhao Huang ,
  • Xiaoman Pan ,
  • Sujian Li ,
  • ,
  • Heng Ji ,
  • Kevin Knight ,
  • Zhen Wen ,
  • Yizhou Sun ,
  • Jiawei Han ,
  • Bulent Yener

ACL-IJCNLP 2015 |

Published by Association for Computational Linguistics (ACL)

DOI | Publication | Publication | Publication

People create morphs, a special type of fake alternative names, to achieve certain communication goals such as expressing strong sentiment or evading censors. For example, “Black Mamba”, the name for a highly venomous snake, is a morph that Kobe Bryant created for himself due to his agility and aggressiveness in playing basketball games. This paper presents the first end-to-end context-aware entity morph decoding system that can automatically identify, disambiguate, verify morph mentions based on specific contexts, and resolve them to target entities. Our approach is based on an absolute “cold-start” it does not require any candidate morph or target entity lists as input, nor any manually constructed morph-target pairs for training. We design a semi-supervised collective inference framework for morph mention extraction, and compare various deep learning based approaches for morph resolution. Our approach achieved significant improvement over the state-of-the-art method (Huang et al., 2013), which used a large amount of training data.