Researcher

Robert C. Moore

Robert C. Moore and Eric Brill published An Improved Error Model for Noisy Channel Spelling Correction at ACL 2000. Moore and Brill were affiliated with Microsoft Research and addressed automatically trained correction of single-word spelling errors. The paper models spelling errors with generic string-to-string edits. It reports improvements over earlier channel models in its experiments, with limits on the error types that the study addresses. Moore and William Lewis’s ACL 2010 paper selects language-model training sentences through a comparison of in-domain and general-model cross-entropy. It reports better models with less training text in the tested setting.

Works
74
Citations
6,093
h-index
32

Publication numbers from OpenAlex, as of 14 October 2026.

Timeline

  1. 2000

    Learning spelling errors

    Robert C. Moore and Eric Brill published An Improved Error Model for Noisy Channel Spelling Correction at ACL 2000. The paper, written at Microsoft Research, addresses automatically trained correction of single-word spelling errors.

  2. 2000 method

    String edits

    The paper models spelling errors with generic string-to-string edits. It reports improvements over earlier channel models in its experiments.

  3. 2010

    Selecting training text

    Moore and William Lewis’s ACL 2010 paper selects language-model training sentences through a comparison of in-domain and general-model cross-entropy. It reports better models with less training text in the tested setting.

Sources and credits

Photo: We have no published photo of Robert C. Moore yet. Send us a free-licensed photo.

  1. An Improved Error Model for Noisy Channel Spelling Correction aclanthology.org. Checked 13 October 2026.
  2. Intelligent Selection of Language Model Training Data aclanthology.org. Checked 13 October 2026.