Information processing system for compaction and replacement of phrases

• US 4,773,039 A
• Filed: 11/19/1985
• Issued: 09/20/1988
• Est. Priority Date: 11/19/1985
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Zamora, Elena M.
• Primary Examiner(s): Shaw, Gareth D.
• Assistant Examiner(s): Chun, Debra A.
• Application Number:
  US06/799,661
• Time in Patent Office: 1,036 Days
• Field of Search: 364/200 MS File, 364/900 MS File, 400/63
• US Class Current: 715/256
• CPC Class Codes: G06F 40/166 Editing, e.g. inserting or deleting
• CPC Class Codes: G06F 40/247 Thesauruses; Synonyms
• CPC Class Codes: G06F 40/253 Grammatical analysis; Style critique
• CPC Class Codes: G06F 40/289 Phrasal analysis, e.g. finite state techniques or chunking
• CPC Class Codes: G06F 40/47 Machine-assisted translation, e.g. using translation memory
• CPC Class Codes: G06F 40/56 Natural language generation

Abstract

An information processing system is disclosed which provides a writer with acceptable replacement phrases to substitute for trite phrases in a manuscript text. The replacement phrases are grammatically equivalent to the trite phrases and can be immediately inserted into the text without further alteration.

Claims

1. In an information processing system including an input unit connected to an input word stream of natural language text, a storage unit for storing natural language text, an execution unit for executing instructions to process natural language text and an output unit for displaying an output word stream of natural language text, a process for the replacement of natural language test, a process for the replacement of natural language source phrases contained in said input word stream with natural language replacement phrases which are inserted into said output word stream, comprising the steps of:

storing a plurality of phrase-pair expressions in said storage unit, each said expression including a source phrase segment containing a variable source word element and a constant source word element and each said expression including a replacement phrase segment containing a variable replacement word element and a constant replacement word element;

storing a source table is said storage unit, having a plurality of source word element values arranged into a plurality of ranks having a grammatically significant sequence, said varaible source word element in said plurality of phrase-pair expressions serving as a pointer for accessing said source table;

storing a replacement table in said storage unit, having a plurality of replacement word element values arranged into a plurality of ranks having a grammatically significant sequence with values in each rank of said replacement table being grammatically equivalent to said values in a corresponding rank of said source table, said variable replacement word element in said phrase-pair expressions serving as a pointer for accessing said replacement table;

comparing in said execution unit, first input words from said input word stream with said constant source word elements in said plurality of phrase-pair expressions and accessing a first one of said phrase-pair expressions having a constant source word element equal to said first input words;

identifying a second input word adjacent to said first input words thus compared;

accessing said source table which is pointed to by said variable source word element in said first one of said phrase-pair expressions, comparing each of said source word element values in said source table with said second input word from said input word stream and identifying the grammatically significant rank of the source word element value in said source table which is equal to said second input word;

accessing said replacement table which is pointed to by said variable replacement word element in said first one of said phrase-pair expressions, and accessing from said replacement table the grammatically equivalent replacement word element value in the rank of said replacement table which corresponds to said grammatically significant rank identified in said source table;

outputting an output replacement phrase to said output unit, which includes said grammatically equivalent replacement word element value and said constant replacement word element from said first one of said phrase-pair expressions.

2. In an information processing system including a means for providing an input word stream of natural language text, a storage means for storing natural language text, an execution means for executing instructions to process natural language text and a utilization means for receiving an output word stream of natural language text, a process for the replacement of natural language source phrases contained in said input word stream with natural language replacement phrases which are inserted into said output word stream, comprising the steps of:

storing a plurality of phrase-pair expressions in said storage means, each said expression including a source phrase segment containing a variable source word element and a constant source word element and each said expression including a replacement phrase segment containing a variable replacement word element and a constant replacement word element;

storing a source table in said storage means, having a plurality of source word element values arranged into a plurality of ranks having a grammatically significant sequence, said variable source word element in said plurality of phrase-pair expressions serving as a pointer for accessing said source table;

storing a replacement table in said storage means, having a plurality of replacement word element values arranged into a plurality of ranks having a grammatically significant sequence with values in each rank of said replacement table being grammatically equivalent to said values in a corresponding rank of said source table, said variable replacement word element in said phrase-pair expressions serving as a pointer for accessing said replacement table;

comparing first input words from said input word stream with said constant source word elements in said plurality of phrase-pair expressions and accessing a first one of said phrase-pair expressions having a constant source word element equal to said first input words;

identifiying a second input word adjacent to said first input words thus compared;

accessing said source table which is pointed to by said variable source word element in said first one of said phrase-pair expressions, comparing each of said source word element values in said source table with said second input word from said input word stream and identifying the grammatically significant rank of the source word element value in said source table which is equivalent to said second input word;

accessing said replacement table which is pointed to by said variable replacement word element in said first one of said phrase-pair expressions, and accessing from said replacement table the grammatically equivalent replacement word element value in the rank of said replacement table which corresponds to said grammatically significant rank identified in said source table;

outputting an output replacement phrase to said utilization means, which includes said grammatically equivalent replacement word element value and said constant replacement word element from said first one of said phrase-pair expressions.

3. The process of claim 2, wherein said step of storing a source table further comprises said plurality of source word element values being a plurality of parts of speech of a verb.

4. The process of claim 2, wherein said step of storing a source table further comprises said plurality of source word element values being a plurality of parts of speech of a pronoun.

5. The process of claim 2, wherein said constant source word element in said phrase-pair expression is hash-encoded, producing a first bit pattern which is stored in a hash screen table.

6. The process of claim 5, wherein said first input words from said input word stream are hash-encoded, producing a second bit pattern which is compared with said first bit pattern in said hash screen table representing said constant source word element.

7. The process of claim 2, wherein said constant source word element includes linguistic selection rules associated therewith for matching words adjacent to said first input words in said input word stream.

8. A process for the compaction and replacement of phrases with grammatically equivalent phrases conforming to conventional grammatical constraints of the original phrase, comprising:

building with a data processor reference lists of pairs of phrases to correspond with each other, for enabling the replacement of a phrase with a grammatically equivalent phrase;

scanning a text to be analyzed, to match a source phrase with a family of target phrases, employing a bit map for source phrases and for target phrases created by hashing a word pair consisting of the least frequently used word and a proximate next-less frequently used word;

continuing with said scanning until a match is obtained in said bit map, indicating that a suitable target phrase has been found;

displaying said target phrase.

9. A process for compiling a phrase table of phrases with grammatically equivalent phrases conforming to conventional grammatical constraints of the original phrase, comprising:

preparing a set of linguistically codified phrases which arranges the order of storage of the elements of the language with which grammatical relationships need to be established and which defines stored tables used to generate conjugations and linguistic variants;

isolating a rarest used word as a file key for each codified phrase;

creating a hash screen for each said file key;

arranging the order of storage phrases in a phrase file by said file key;

generating characteristic match terms from phrases and compacting said set of codified phrases based upon character frequency;

storing said match terms for said compacted phrases and hash screen for reference by linguistic decoder.

10. A process for decoding phrase tables for replacing phrases with grammatically equivalent phrases conforming to conventional grammatical constraints of the original phrase, comprising the steps of:

scanning an input text containing a plurality of phrases;

hashing with a data processor a word selected from said plurality of input phrases;

comparing said hashed word with a hash bit screen for a phrase file of phrases equivalent to said selected word;

matching adjacent words to said selected word based upon linguistic rules relating said input phrase with said phrase derived from said phrase file;

outputting a replacement phrase from said phrase file which satisfies said matching step.