System for analyzing translations
First Claim
1. Translation analysis apparatus, the apparatus operating on first and second integral unaligned texts where the second text is at least in part a translation of at least a portion of a first text, the apparatus comprising:
- means for storing a set of pairs of locations, each pair of locations including a first location in the first integral unaligned text and a second location in the second integral unaligned text that is in the neighborhood of the translation of the first location, the set of pairs of location being markers;
means for responding to a specification of a given location in the first text by employing positioning from the set of pairs of locations to determine a specification of a corresponding location in the second text that is in the neighborhood of the translation of the given location in the first location, by using the positioning from the set of pairs of locations; and
concordance making means for finding a first set of locations of a term in the first text, making a second set of the corresponding locations in the neighborhood of a corresponding term translation in the second text by providing the locations in the first set thereof to the responding means, and making the concordance using the term location in the first set and the corresponding term location in the second set.
6 Assignments
0 Petitions
Accused Products
Abstract
Apparatus and methods for comparing a text with its translation. The apparatus uses cognates in the text and the translation to make a list of first positions in the text and second positions in the translation. The second position specifies a location in the translation which is in the neighborhood of the translation of the first position in the text. By means of the list, a given location in the text can be related to a location in the translation which is in the neighborhood of the translation of the given location. The list permits parallel displays of the text and the translation and easy production of full-context concordances, glossaries, and histograms. Graphs based on the list can be used to detect parts of the text which are not in the translation and vice-versa. The list is made using a technique which iteratively computes an alignment path in a graph which records the positions of matching 4-grams in the text and the translation. The technique is more robust than techniques which use sentences and paragraphs to determine alignment.
-
Citations
22 Claims
-
1. Translation analysis apparatus, the apparatus operating on first and second integral unaligned texts where the second text is at least in part a translation of at least a portion of a first text, the apparatus comprising:
-
means for storing a set of pairs of locations, each pair of locations including a first location in the first integral unaligned text and a second location in the second integral unaligned text that is in the neighborhood of the translation of the first location, the set of pairs of location being markers; means for responding to a specification of a given location in the first text by employing positioning from the set of pairs of locations to determine a specification of a corresponding location in the second text that is in the neighborhood of the translation of the given location in the first location, by using the positioning from the set of pairs of locations; and concordance making means for finding a first set of locations of a term in the first text, making a second set of the corresponding locations in the neighborhood of a corresponding term translation in the second text by providing the locations in the first set thereof to the responding means, and making the concordance using the term location in the first set and the corresponding term location in the second set. - View Dependent Claims (2, 3, 4, 8, 9)
-
-
5. Translation analysis apparatus, the apparatus operating on first and second texts where the second text is at least in part a translation of at least a portion of a first text, the apparatus comprising:
-
means for storing a set of pairs of locations, each pair of locations including a first location in the first text and a second location in the second text that is in the neighborhood of the translation of the first location; means for producing the set of pairs of locations by making a graph of the occurrences of markers in the first text and the second text, determining an alignment path in the graph, and producing the set of pairs of locations from the alignment path; means for responding to a specification of a given location in the first text by employing the set of pairs of locations to determine a specification of a corresponding location in the second text which is in the neighborhood of the translation of the first location; and histogram making means for finding a first set of locations of a term in the first text, making a second set of the corresponding locations by providing the locations in the first set thereof to the means for responding to a specification of a given location, and making the histogram using the locations in the second set, the histogram showing frequencies of occurrences of words in the second set of locations.
-
-
6. Translation analysis apparatus, the apparatus operating on first and second texts where the second text is at least in part a translation of at least a portion of a first text, the apparatus comprising:
-
means for storing a set of pairs of locations, each pair of locations including a first location in the first text and a second location in the second text which is in the neighborhood of the translation of the first location; means for responding to a specification of a given location in the first text by employing the set of pairs of locations to determine a specification of a corresponding location in the second text which is in the neighborhood of the translation of the first location; and means for detecting differences between the first text and the second text by making a graph from the set of pairs of locations which shows a relationship between the pairs of locations in the set and other parts of the texts and analyzing the graph by comparing a curve made from the set of pairs of locations with a predetermined approximation thereof. - View Dependent Claims (7)
-
-
10. Translation analysis apparatus, the apparatus operating on first and second texts where the second text is at least in part a translation of at least a portion of a first text, the apparatus comprising:
-
means for making a set of pairs of locations by comparing letters in the first text with letters in the second text, each pair of locations including a first location in the first text and a second location in the second text which is in the neighborhood of the translation of the first location; means for producing the set of pairs of locations by making a graph of the occurrences of cognates in the first text and the second text, determining an alignment path in the graph, and using the alignment path to produce the set of pairs of locations; means for storing the set of pairs of locations; and means for responding to a specification of a given location in the first text by employing the set of pairs of locations to determine a specification of a corresponding location in the second text which is in the neighborhood of the translation of the first location. - View Dependent Claims (11, 12, 13, 14, 15)
-
-
16. A method employed with a first text and a second text, the second text being at least in part a translation of at least a portion of the first text, to make a set of pairs of locations stored in a computer system, each pair of locations including a first location in the first text and a second location in the second text which is in the neighborhood of the translation of the first location, the method comprising the steps performed in the computer system of:
-
using a representation of occurrences of cognates in the first text and the second text stored in the computer system and processing means in the computer system to determine an alignment path; and using the processing means to derive the pairs of locations from the alignment path.
-
-
17. A method employed with a first text and a second text for determining an alignment path for the first text and the second text, the method comprising the steps performed in a computer system of:
-
employing processing means in the computer system to make an F-image and store the F-image in the computer system, the F-image having a plurality of cells, each cell representing a first portion of the first text and a second portion of the second text and each cell being given a first value if the first portion and the second portion do not contain the same n-gram and a second value if they do; and employing the processing means to compute the alignment path from the F-image. - View Dependent Claims (18, 19, 20, 21, 22)
-
Specification