An Improved Error Model for Noisy Channel Spelling Correction

By Eric Brill, Robert C. Moore. 2000

632 citations recorded by the source corpus.

The noisy channel model has been applied to a wide range of problems, including spelling correction. These models consist of two components: a source model and a channel model. Very little research has gone into improving the channel model for spelling correction. This paper describes a new channel model for spelling correction, based on generic string to string edits. Using this model gives significant performance improvements compared to previously proposed models.

Subfields: Artificial Intelligence, Hardware and Architecture, Natural Language Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Hardware and Architecture

In Natural Language Processing

Citations in the corpus

BPE-Dropout: Simple and Effective Subword Regularization (2019, methodology): “Edit distance is commonly used to model misspellings (Brill and Moore, 2000; Ahmad and Kondrak, 2005; Pinter et al., 2017).”
The Neural Noisy Channel (2016, background): “Noisy channel decompositions have been successfully used in a variety of problems, including speech recognition (Jelinek, 1998), machine translation (Brown et al., 1993), spelling correction (Brill & Moore, 2000), and question answering (Echihabi & Marcu, 2003).”
A Large Scale Ranker-Based System for Search Query Spelling Correction (2010, methodology): “The former focuses on the development of error models based on different edit distance functions (e.g., Kucich, 1992; Kernighan et al., 1990; Brill and Moore, 2000; Toutanova and Moore, 2002).”
Revisiting Challenges in Data-to-Text Generation with Fact Grounding (2020, neutral): “= |cp′| – Precision (RGP) = |cp′ ∩ S| / |cp′| • Content Ordering (CO): – DLD: normalized Damerau Levenshtein Distance (Brill and Moore, 2000) between cp and cp′ CS and RG measures the “what to say” and CO measures the “how to say” aspects.”
Sequence Synopsis: Optimize Visual Summary of Temporal Event Data (2018, methodology): “In the algorithm, the minimum number of edits can be determined by following the method to calculate the Damerau-Levenshtein distance [2].”
Combating Adversarial Misspellings with Robust Word Recognition (2019, methodology): “Classic methods rely on a source language model and a noisy channel model to find the most likely correction for a given word (Mays et al., 1991; Brill and Moore, 2000).”
Learning a Spelling Error Model from Search Query Logs (2005, background): “In (Brill and Moore, 2000), a more general model allowing generic string to string ed-its is used, allowing many-to-one and one-to-many character substitution edits.”
Lexical Normalisation of Short Text Messages: Makn Sens a #twitter (2011, background): “Brill and Moore (2000) characterise the error model by computing the product of operation probabilities on slice-by-slice string edits.”
Moral Stories: Situated Reasoning about Norms, Intents, Actions, and their Consequences (2020, methodology): “Splits are obtained by ordering stories according to the Damerau–Levenshtein distance (DL) (Brill and Moore, 2000) between their actions (or consequences) and assigning stories with lowest distances to the test set, followed by the development set.”
Dealing with Input Noise in Statistical Machine Translation (2012, background): “Generally, misspelling problems can be addressed with a simple Levenshtein distance under a noisy channel model paradigm (Brill and Moore, 2000).”
Building Language Models for Text with Named Entities (2018, background): “Language model is a fundamental component in Natural Language Processing (NLP) and it supports various applications, including document generation (Wiseman et al., 2017), text auto-completion (Arnold et al., 2017), spelling correction (Brill and Moore, 2000), and many others.”
QuickEdit: Editing Text & Translations via Simple Delete Actions (2017, methodology): “In particular, statistical models are used for that purpose, for instance to help correct spelling mistakes (Brill and Moore, 2000) or suggest likely completions of a sentence (Bickel et al., 2005).”