Automatic detection and application of editing patterns in draft documents

• US 8,086,453 B2
• Filed: 01/26/2009
• Issued: 12/27/2011
• Est. Priority Date: 11/08/2005
• Status: Active Grant

• Current Assignee: Solventum Intellectual Properties Company (Solventum Corp.)
• Original Assignee: Multimodal Technologies Incorporated (3M Company)
• Inventors: Detlef, Koll, Fritsch, Juergen, Finke, Michael
• Primary Examiner(s): Armstrong, Angela A
• Application Number:
  US12/360,109
• Publication Number:
  US 20090132911A1
• Time in Patent Office: 1,065 Days
• Field of Search: 704/9, 704/235, 704/270
• US Class Current: 704/235
• CPC Class Codes: G06F 40/232 Orthographic correction, e.g. spell checking or vowelisation
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence

Abstract

An error detection and correction system extracts editing patterns and derives correction rules from them by observing differences between draft documents and corresponding edited documents, and/or by observing editing operations performed on the draft documents to produce the edited documents. The system develops classifiers that partition the space of all possible contexts into equivalence classes and assigns one or more correction rules to each such class). Once the system has been trained, it may be used to detect and (optionally) correct errors in new draft documents. When presented with a draft document, the system identifies first content (e.g., text) in the draft document and identifies a context of the first content. The system identifies a correction rule based on the first content and the first context. The system may use a classifier to identify the correction rule. The system applies the correction rule to the first content to produce second content.

Claims

1. A method performed by a computer processor executing computer program instructions tangibly stored on a non-transitory computer readable medium, the method comprising:

(A) identifying a plurality of editing patterns of the form T=(D,E,C), wherein each of the plurality of editing patterns relates content D in an original document corpus to corresponding content E in an edited document corpus in a context C shared by contents D and E; and

(B) identifying a plurality of likelihoods of correctness of the plurality of editing patterns, wherein (B) comprises;

(B) (1) counting a number of positive instances in which content D in context C in the original document corpus has been replaced with content E in context C in the edited document corpus, and identifying a likelihood of correctness of the editing pattern T=(D,E,C) based on the number of positive instances; and

(B) (2) counting a number of negative instances in which content D in context C in the original document corpus remained unchanged in context C in the edited document corpus, and identifying a likelihood of correctness of the editing pattern T=(D,E,C) based on the number of positive instances and the number of negative instances;

(C) selecting one of the plurality of editing patterns based on the plurality of likelihoods of correctness; and

(D) applying the selected editing pattern to replace an instance of content D in a further document with an instance of content E in the further document.

2. The method of claim 1, wherein (B) comprises counting the number of instances by comparing documents in the original document corpus to documents in the edited document corpus to identify differences between them.

3. The method of claim 1, wherein (B) comprises identifying editing actions performed on the original document corpus to replace content D in context C in the original document corpus to produce content E in context C in the edited document corpus.

4. The method of claim 1, wherein (C) comprises selecting the maximum likelihood of correctness among the plurality of likelihoods of correctness.

5. A non-transitory computer readable medium comprising computer program instructions executable by a computer processor to perform a method, the method comprising:

(A) identifying a plurality of editing patterns of the form T=(D,E,C), wherein each of the plurality of editing patterns relates content D in an original document corpus to corresponding content E in an edited document corpus in a context C shared by contents D and E; and

(B) identifying a plurality of likelihoods of correctness of the plurality of editing patterns, wherein (B) comprises;

(B) (1) counting a number of positive instances in which content D in context C in the original document corpus has been replaced with content E in context C in the edited document corpus, and identifying a likelihood of correctness of the editing pattern T=(D,E,C) based on the number of positive instances; and

(B) (2) counting a number of negative instances in which content D in context C in the original document corpus remained unchanged in context C in the edited document corpus, and identifying a likelihood of correctness of the editing pattern T=(D,E,C) based on the number of positive instances and the number of negative instances;

(C) selecting one of the plurality of editing patterns based on the plurality of likelihoods of correctness; and

(D) applying the selected editing pattern to replace an instance of content D in a further document with an instance of content E in the further document.

6. The computer readable medium of claim 5, wherein (B) comprises counting the number of instances by comparing documents in the original document corpus to documents in the edited document corpus to identify differences between them.

7. The computer readable medium of claim 5, wherein (B) comprises identifying editing actions performed on the original document corpus to replace content D in context C in the original document corpus to produce content E in context C in the edited document corpus.

8. The computer readable medium of claim 5, wherein (C) comprises selecting the maximum likelihood of correctness among the plurality of likelihoods of correctness.

9. A method performed by a computer processor executing computer program instructions tangibly stored on a non-transitory computer readable medium, the method comprising:

(A) identifying a plurality of editing patterns of the form T=(D,E,C), wherein each of the plurality of editing patterns relates content D in an original document corpus to corresponding content E in an edited document corpus in a context C shared by contents D and E; and

(B) identifying a plurality of likelihoods of correctness of the plurality of editing patterns;

(C) receiving an input from a user representing a selection of one of the plurality of editing patterns;

(D) applying the selected editing pattern to replace an instance of content D in a further document with an instance of content E in the further document; and

(E) increasing the likelihood of correctness of the selected editing pattern in response to the selection;

wherein the plurality of likelihoods of correctness includes an initial likelihood of correctness of the selected editing pattern;

wherein the selected editing pattern comprises a positive editing pattern;

wherein the initial likelihood of correctness of the selected editing pattern is based on a number of occurrences of the positive editing pattern; and

wherein (E) comprises incrementing the number of occurrences of the positive editing pattern and increasing the likelihood of correctness of the selected editing pattern based on the incremented number of occurrences of the positive editing pattern.

10. The method of claim 9:

wherein the initial likelihood of correctness of the selected editing pattern is based on a number of occurrences of the positive editing pattern and a number of occurrences of a negative editing pattern corresponding to the positive editing pattern.

11. The method of claim 9, wherein (C) further comprises determining that none of the plurality of editing patterns satisfies predetermined decision criteria, and prompting the user for the input representing the selection of one of the plurality of editing patterns in response to the determination.

12. The method of claim 9, wherein (C) further comprises prompting the user with output representing at least some of the plurality of editing patterns presented in a descending order corresponding to the likelihoods of correctness of the at least some of the plurality of editing patterns.

13. A non-transitory computer readable medium comprising computer program instructions executable by a computer processor to perform a method, the method comprising:

(A) identifying a plurality of editing patterns of the form T=(D,E,C), wherein each of the plurality of editing patterns relates content D in an original document corpus to corresponding content E in an edited document corpus in a context C shared by contents D and E; and

(B) identifying a plurality of likelihoods of correctness of the plurality of editing patterns;

(C) receiving an input from a user representing a selection of one of the plurality of editing patterns;

(D) applying the selected editing pattern to replace an instance of content D in a further document with an instance of content E in the further document; and

(E) increasing the likelihood of correctness of the selected editing pattern in response to the selection;

wherein the plurality of likelihoods of correctness includes an initial likelihood of correctness of the selected editing pattern;

wherein the selected editing pattern comprises a positive editing pattern;

wherein the initial likelihood of correctness of the selected editing pattern is based on a number of occurrences of the positive editing pattern; and

wherein (E) comprises incrementing the number of occurrences of the positive editing pattern and increasing the likelihood of correctness of the selected editing pattern based on the incremented number of occurrences of the positive editing pattern.

14. The computer readable medium of claim 13:

wherein the initial likelihood of correctness of the selected editing pattern is based on a number of occurrences of the positive editing pattern and a number of occurrences of a negative editing pattern corresponding to the positive editing pattern.

15. The computer readable medium of claim 13, wherein (C) further comprises determining that none of the plurality of editing patterns satisfies predetermined decision criteria, and prompting the user for the input representing the selection of one of the plurality of editing patterns in response to the determination.

16. The computer readable medium of claim 13, wherein (C) further comprises prompting the user with output representing at least some of the plurality of editing patterns presented in a descending order corresponding to the likelihoods of correctness of the at least some of the plurality of editing patterns.