We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.CL

Change to browse by:

cs

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computation and Language

Title: End-to-end Neural Coreference Resolution Revisited: A Simple yet Effective Baseline

Abstract: Since the first end-to-end neural coreference resolution model was introduced, many extensions to the model have been proposed, ranging from using higher-order inference to directly optimizing evaluation metrics using reinforcement learning. Despite improving the coreference resolution performance by a large margin, these extensions add a lot of extra complexity to the original model. Motivated by this observation and the recent advances in pre-trained Transformer language models, we propose a simple yet effective baseline for coreference resolution. Our model is a simplified version of the original neural coreference resolution model, however, it achieves impressive performance, outperforming all recent extended works on the public English OntoNotes benchmark. Our work provides evidence for the necessity of carefully justifying the complexity of existing or newly proposed models, as introducing a conceptual or practical simplification to an existing model can still yield competitive results.
Subjects: Computation and Language (cs.CL)
Cite as: arXiv:2107.01700 [cs.CL]
  (or arXiv:2107.01700v2 [cs.CL] for this version)

Submission history

From: Tuan Manh Lai [view email]
[v1] Sun, 4 Jul 2021 18:12:24 GMT (126kb,D)
[v2] Tue, 13 Jul 2021 06:32:09 GMT (127kb,D)

Link back to: arXiv, form interface, contact.