Natural Language Understanding Wiki

From Rahman & Ng (2012)[1]: "We employ narrative chains to heuristically predict the antecedent for the target pronoun, and encode the prediction as a feature. [...] we extract all the narrative chains containing both elements in the pair from Chambers and Jurafsky’s output."

From Khashabi & Roth (2015)[2] propose "Predicate schemas" and use them to solve Winnograd-style and regular coreference resolution tasks.

"The first specifies one predicate with its subject and object, thus providing information on the subject and object preferences of a given predicate. The second specifies two predicates with a semantically shared argument (either subject or object), thus specifies role preferences of one predicate, among roles of the other. We instantiate these schemas by acquiring statistics in an unsupervised way from multiple resources including the Gigaword corpus, Wikipedia, Web Queries and polarity information."


Pichotta (2015)[3] propose to do the same (see photo) but the work is not done yet.

Peng and Roth (2016)[4] used discourse-driven language model (also semantics model?) to improve coreference resolution.

See also[]


  1. Rahman, A., & Ng, V. (2012). Resolving Complex Cases of Definite Pronouns : The Winograd Schema Challenge. Proceedings of the 2012 Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning, (July), 777–789.
  2. Peng, H., Khashabi, D., & Roth, D. (2015). Solving Hard Coreference Problems. Proc. of NAA, 61801, 809–819. Retrieved from
  3. Pichotta, K. (2015). Statistical Script Learning with Recurrent Neural Nets, 1–40.
  4. Haoruo Peng and Dan Roth. 2016. Two Discourse Driven Language Models for Semantics. ACL.