Systems and methods for user-interest sensitive note-taking

• US 7,827,029 B2
• Filed: 11/30/2004
• Issued: 11/02/2010
• Est. Priority Date: 11/30/2004
• Status: Expired due to Fees

• Current Assignee: Palo Alto Research Center, Inc. (Xerox Holdings Corp.)
• Original Assignee: Palo Alto Research Center, Inc. (Xerox Holdings Corp.)
• Inventors: Tepper, Michael, Bobrow, Daniel, Kaplan, Ronald, Crouch, Richard
• Primary Examiner(s): Hudspeth; David R
• Assistant Examiner(s): Spooner; Lamont M
• Application Number:
  US10/999,793
• Publication Number:
  US 20060116861A1
• Time in Patent Office: 2,163 Days
• Field of Search: 704/9, 704/1, 704/10, 707 1- 6, 707/736, 707/767, 715/254
• US Class Current: 704/9
• CPC Class Codes: G06F 40/258 Heading extraction; Automatic titling; Numbering
• CPC Class Codes: G06F 40/30 Semantic analysis

Abstract

Techniques are presented to determine user-interest sensitive notes. User selected passages, user interest information, condensation transformations and optional meaning distortion constraints are identified. User foci expressed by the selected passages are determined based on the similarity of the elements in the selected passages to elements in the user interest information. User sensitive notes are determined by selectively applying the condensation transformations to the selected passages to preferentially retain user foci while eliding less salient information. Meaning distortions constraints are optionally applied in conjunction with condensation transformations or in creating the condensation transformations in order to reduce the likelihood of distorting the meaning of the passage.

Claims

1. A method comprising the steps of:

determining a user selected passage by highlighting, marking, or cut and pasting a block of text using a computer operable input device;

determining user interest information;

determining a parsing grammar and a generation grammar;

determining a meaning structure based on the selected passage;

transforming the user selected passage and the user interest information into condensation transformations comprised of rewrite rules which delete, merge and change elements of the meaning structure;

determining meaning distortion constraints beyond which the meaning structure is identified as being distorted after the condensation transformations are applied;

determining a packed meaning structure based on the selected passage and the determined parsing grammar;

determining a reduced meaning structure based on the condensation transformations, the packed meaning structure, the user information, and the meaning distortion constraints;

determining at least one candidate condensation structure based on the reduced meaning structure and a predictive model or a statistical disambiguation model; and

determining a user-interest note based on the candidate condensation structures and the determined generation grammar.

2. The method of claim 1, in which determining the note further comprises the step of prioritizing the retention of the elements of the passage that match elements of the user interest information.

3. The method of claim 2, in which the matches are determined based on at least one of:

literal, distributional, statistical, and co-occurrence similarity.

4. The method of claim 3, in which the conceptual similarity is based on at least one of:

a lexicon;

an ontology;

statistics and a machine-readable information source.

5. The method of claim 1, in which the meaning structure is determined based on a parsing grammar.

6. The method of claim 5, in which the parsing grammar is determined based on at least one of:

a characteristic of the passage and user preference.

7. The method of claim 6, in which the characteristic of the passage is at least one of:

a language and a genre characteristic.

8. The method of claim 1, in which the note is comprised of grammatical sentences.

9. The method of claim 1, in which the passage is at least one sentence.

10. The method of claim 9, in which the at least one sentence is comprised of at least two words.

11. The method of claim 1, in which the packed meaning structure is at least one of:

a logical formula in the predicate calculus, a collection of semantic facts, a head-driven phrase structure and an f-structure.

12. The method of claim 1, in which the passage is converted into meaning structures using a parsing grammar.

13. The method of claim 1, in which the condensation rules are re-write rules.

14. The method of claim 1, in which the condensation transformations are applied to the meaning structures based on meaning distortion constraints.

15. The method of claim 14, in which the meaning distortion constraints are useable to adjust the condensation of meaning structures characterized as downward monotonic.

16. The method of claim 15, in which the meaning distortion constraint adjustment is at least one of:

preventing condensation and repairing meaning in the meaning structure.

17. The method of claim 1, in which the determined note is comprised of a portion that signals the meaning of the passage.

18. The user-interest sensitive note-viewing system as determined by the method of claim 1 comprising:

a computer operable memory;

an input/output circuit that receives user input and retrieves stored user-interest sensitive notes from memory; and

a processor that displays the user-interest sensitive note, wherein the user-interest sensitive note can edited by at least one user.

19. The system of claim 18, in which the input/output circuit interprets user-interest information and the processor dynamically updates the condensate portion of the user-interest sensitive note based on the retrieved user-interest information.

20. The system of claim 19, in which the user interest sensitive note is comprised of a condensate portion.

21. The user interest sensitive note as determined by the method of claim 1, in which the user-interest sensitive note is comprised of at least one condensate and in which a user selected link provides access to the passage.

22. A system for generating a user-interest note comprising:

a processor that selects a text passage by highlighting the text or by inputting the text;

user interest information received from at least one of entered keywords, monitored internet browsing patterns, and user academic or professional information;

at least one condensation transformation expressed as rewrite rules and functions which reduces the selected passage by merging, deleting and changing the elements of the selected passage to remove modifiers and less salient information;

a meaning structure determined by using a parsing grammar;

a reduced meaning structure derived from the selected passage and based on input user information;

meaning distortion constraints beyond which the meaning structure is identified as being distorted after the at least one condensation transformation is applied; and

a plurality of candidate condensate based on the reduced meaning structure which are optimally ranked based on a predictive model or a statistical disambiguation model;

wherein the processor further selects a user-interest note from the candidate condensates based on the input user interests.

23. The system of claim 22, in which elements of the passage that match elements of the user interest information are preferentially retained.

24. The system of claim 23, in which the matches are determined based on at least one of:

literal, conceptual, distributional, statistical, and co-occurrence similarity.

25. The system of claim 24, in which the conceptual similarity is based on a lexicon;

an ontology;

statistics and a machine-readable information source.

26. The system of claim 22, in which the processor determines a meaning structure for the passage;

and further determines the condensate based on elements of the meaning structure that match elements of the user interest information.

27. The system of claim 26, in which the meaning structure is determined based on a parsing grammar.

28. The system of claim 27, in which the parsing grammar is determined based on at least one of:

a characteristic of the passage and user preference.

29. The system of claim 28, in which the characteristic of the passage is at least one of:

a language and a genre characteristic.

30. The system of claim 26, in which the meaning structure is a packed meaning structure.

31. The system of claim 30, in which the packed meaning structure is at least one of:

a logical formula in the predicate calculus, a collection of semantic facts, a head-driven phrase structure and an f-structure.

32. The system of claim 22, in which the condensate is comprised of grammatical sentences.

33. The system of claim 22, in which the passage is at least one sentence.

34. The system of claim 33, in which the at least one sentence is comprised of at least two words.

35. The system of claim 22, in which the passage is converted into at least one meaning structure using a parsing grammar.

36. The system of claim 22, in which the condensation rules are re-write rules.

37. The system of claim 22, in which the passage is converted into a packed meaning structure and the condensation transformations are applied to the packed structure based on meaning distortion constraints.

38. The system of claim 37, in which the meaning distortion constraints adjust the condensation of the meaning structures characterized as downward monotonic.

39. The system of claim 38, in which the meaning distortion constraint adjustment is at least one of:

preventing condensation and adjusting the meaning in the meaning structure.

40. A method of creating a user-interest note comprising the steps of:

determining a passage by highlighting text or by entering text;

determining user interest information by at least one of entered keywords, monitored Internet browsing patterns, and user academic or professional information;

determining condensation transformations expressed as rewrite rules and functions;

determining at least one reduced passage based on user interest information;

determining meaning distortion constraints beyond which the at least one reduced passage is identified as being distorted after the condensation transformations are applied;

determining at least one text condensate based on comparison between the reduced passage, the user interest information, and the condensation transformations; and

selecting one user-interest note from the at least one reduced passage.

41. A method of using a computer to generate a user-interest note, the method comprising:

selecting a text input passage by use of a computer to highlight text or input text;

parsing the input passage into elements using a computer processor;

determining user foci based on similarities between the user information and the elements of the selected text input passage;

determining a condensation transformation expressed as rewrite rules and functions which reduce the selected passage by merging, deleting, and changing the elements of the selected passage;

determining a reduced passage text note by removing information from the passage based on the condensation transformation, the user foci, and the user interest information;

and

determining meaning distortion constraints beyond which the reduced passage text condensation is identified as being distorted after the condensation transformation is applied to the passage.

42. A computer readable storage medium comprising computer readable program code embodied on the computer readable storage medium, the computer readable program code usable to program a computer for generating a user-interest note comprising the steps of:

determining a passage by highlighting one or more multi-word sentences through use of a computer mouse controlled cursor wherein a mouse button is held down as the cursor is dragged over a selected passage in order to capture the passage;

determining user interest information by parsing the text of the user selected passage to determine alternative meaning structures associated with the passage;

determining condensation transformations expressed as rewrite rules and functions which reduce the selected passage by deleting, merging, and/or adjusting the element of alternative meaning structures in order to compare user interest information with alternative meaning structures;

determining meaning distortion constraints beyond which the alternative meaning structures are identified as being distorted after the condensation transformations are applied; and

applying a condensation transformation based on the determined passage, and the user interest information to condense facts in the alternative meaning structures to create condensed packed meaning structure.

43. A user-interest note derived by using a computer processor to transform a user selected passage of a text into a packed meaning structure which is at least one of:

a logical formula in the predicate calculus,

a minimal recursion semantic structure,

a head-driven phrase structure, and

an f-structure; and

wherein the user interest note and

is based on one or more input indicators of user interest and includes;

a condensation transformation expressed as rewrite rules and functions which reduce the selected passage by merging, deleting, and changing elements of the selected passage;

a reduced passage text note by removing information from the passage based on the condensation transformation, the user interest, and user interest information; and

meaning distortion constraints beyond which the reduced passage text condensation is identified as being distorted after the condensation transformation is applied to the passage,

wherein the meaning structure for a passage is created through the use of a stochastic disambiguation model and/or predictive model which assigns higher probabilities to better examples from a training set and assigns lower probabilities to less desirable or less appropriate examples.

44. The note of claim 43, further comprising at least one selectable user interface component operable to adjust a display of a portion of the user-interest note.

45. The note of claim 44, wherein the displayed portion of the user-interest sensitive note is at least one of:

the passage portion and the condensate portion.

46. The note claim 44, wherein the user selectable interface component is at least one of:

a button, a link and a uniform resource locator.

47. The note of claim 43, wherein the condensate contains information from the passage portion.

48. The note of claim 43, wherein the passage portion is displayed in at least one of:

a window;

a dialog box;

a pop-up-box and a status area.

49. The note of claim 44, wherein the selectable user interface component is responsive to at least one of:

voice, touch and eye-gaze input.

50. A method of creating a corrected user-interest note comprising the steps of:

determining a user selected text passage by highlighting or inputting one or more multi-word sentences;

parsing the user selected text into elements using a parsing grammar;

determining user interest information from at least one of entered keywords, monitored Internet browsing patterns, user academic information and user professional information;

determining condensation transformations expressed as rewrite rules and functions;

determining non-local references in the passage;

determining a note based on comparison between the determined passage, the user interest information, and the condensation transformations;

determining meaning distortion constraints beyond which the note is identified as being distorted after the application of at least one condensation transformation;

identifying dropped words and meaning distortions within the note; and

producing a corrected note based on the text condensation and the addition of dropped words and repairing of meaning distortions.