Multi-dimensional method and apparatus for automated language interpretation
First Claim
1. A schema for storing, in digital, analog, or other machine-readable format, a proposition formed of a plurality of objects, comprising:
- a molecular-style structure having a plurality of objects wherein said objects function as elements x, y, r, and c and wherein said objects form an explicit proposition comprising at least one of each x, y, r, and c element, wherein x is a first object, y is a second object, r represents a known explicit relationship between said first and second objects, c represents a known logical context of said relationship;
said relationship forming a logical proposition stating a fact regarding said first object; and
said proposition having a predetermined weight assigned thereto.
1 Assignment
0 Petitions
Accused Products
Abstract
A method and apparatus for natural language interpretation are described. The invention includes a schema and apparatus for storing, in digital, analog, or other machine-readable format, a network of propositions formed of a plurality of text and/or non-text objects, and the steps of retrieving a string of input text, and locating all associated propositions in the network for each word in the input string. Embodiments of the invention also include optimization steps for locating said propositions, and specialized structures for storing them in a ready access storage area simulating human short-term memory. The schema and steps may also include structures and processes for obtaining and adjusting the weights of said propositions to determine posterior probabilities representing the intended meaning. Embodiments of the invention also include an apparatus designed to apply an automated interpretation algorithm to automated voice response systems and portable knowledge appliance devices.
-
Citations
163 Claims
-
1. A schema for storing, in digital, analog, or other machine-readable format, a proposition formed of a plurality of objects, comprising:
-
a molecular-style structure having a plurality of objects wherein said objects function as elements x, y, r, and c and wherein said objects form an explicit proposition comprising at least one of each x, y, r, and c element, wherein x is a first object, y is a second object, r represents a known explicit relationship between said first and second objects, c represents a known logical context of said relationship;
said relationship forming a logical proposition stating a fact regarding said first object; and
said proposition having a predetermined weight assigned thereto. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16)
-
-
17. A schema for storing, in digital, analog, or other machine-readable format, a plurality of propositions formed of a plurality of text or non-text objects, comprising:
-
a storage architecture having a plurality of objects therein, each said object being linked within said storage architecture to another said object, said links defined by the relationship of said two objects in the context of said formal logic proposition; and
an object network structure having an arbitrarily large plurality of molecular groupings of objects in which each molecular grouping of objects represents a formal logic proposition and each said object molecule is formed of a plurality of objects, each with a specified role in forming said logical proposition. - View Dependent Claims (18, 19, 20, 21, 22)
-
-
23. A storage architecture for storing and processing proposition information in digital, analog, and other machine-readable formats, comprising;
-
a permanent storage area for retaining a body of proposition information therein;
a working storage area for temporary storage and analysis of portions of said information; and
a ready access storage area for retaining a portion of said information from said working storage area, said information in said working storage area being copied from at least one of said other storage areas. - View Dependent Claims (24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45)
-
-
46. A method for the retrieval of data in an object network structure, comprising the steps of;
-
providing a computer architecture having a object base with a plurality of objects therein, each object representing a letter, number, word, symbol, or data file and each object having a set of propositions linked thereto, said propositions describing the relationships of a first object to at least one other object and said set of propositions accessible by location of said first object;
searching said object base for a predetermined first object; and
accessing said set of propositions.
-
-
47. A method for the analysis of the meaning of an ordered group of objects, comprising the steps of;
-
providing a plurality of propositions formed of two or more objects, each said proposition describing an explicit logical relationship between one object and another object;
comparing all propositions containing a first object of said group of objects to identify any common objects; and
assigning a weight to each said proposition based upon the amount of common objects identified in said proposition. - View Dependent Claims (48, 49, 50)
-
-
51. A method for the analysis of the meaning of an ordered group of objects, comprising the steps of;
-
providing a plurality of propositions ordered sequentially and formed of a plurality of objects, each said proposition describing an explicit logical relationship between one object and another object and having objects representing x, y, r, and c, wherein x is a first object, y is a second object, r is an explicit logical relationship between said first object x and second object y, and c is a logic context in which said logical relationship exists between said first object x and said second object y, said propositions having said objects representing x, y, r, and c ordered respectively;
comparing all propositions containing a first object of said group of objects to identify any common objects;
sequentially examining said propositions for common objects;
extracting y objects from said propositions and reordering them in said proposition as x objects for further sequential examination;
sequentially examining said reordered propositions for common objects;
extracting r relations from said propositions to serve to activate further steps specified by the type or specific identity of the r relation;
sequentially examining said propositions for common objects;
extracting c contexts from said propositions to categorize or group the propositions with other propositions sharing the same or similar c contexts;
sequentially examining said propositions for common objects;
comparing all propositions containing a first object of said group of objects to identify any common objects;
assigning a weight to each said proposition based upon the amount of common objects or the frequency of encountering common objects during sequential examination.
-
-
52. A storage architecture for storing digital, analog, or other machine-readable format material, comprising;
-
an object base having a plurality of objects therein; and
a set of propositions linked to a first object, said set of propositions describing the relationships of a first object to at least one other object, wherein each proposition describes the relationship between said first object and at least one other object. - View Dependent Claims (53, 54, 55, 56, 57, 58, 59)
-
-
60. A method of for the analysis of the meaning of an ordered group of objects, comprising the steps of;
-
providing a storage architecture having an object base with a plurality of objects therein;
providing a plurality of propositions formed of a plurality objects, each said proposition describing an explicit logical relationship between one object and another object;
matching a first object, in said group of objects, with an identical object in said object base;
identifying a number of propositions containing said first object; and
examining said propositions for objects common between at least one of said propositions and said ordered group of objects. - View Dependent Claims (61, 62, 63)
-
-
64. A method of selecting at least one proposition from a plurality of propositions, the steps of the method comprising:
-
providing a plurality of propositions, formed from a plurality of objects, each said proposition having a predetermined weight and vector, said weight having a numeric value and said vector having a direction and magnitude;
selecting at least one proposition from said plurality of propositions; and
evaluating said selected propositions based upon said weight values and direction and magnitude of said vectors. - View Dependent Claims (65, 66, 67, 68)
-
-
69. A method for interpreting text, formed of a plurality of words, comprising the steps of:
-
providing a lexicon having a plurality of words therein, each said word being linked to a plurality of propositions formed by a plurality of words with each said linked proposition having said words to which said proposition is linked forming a portion of said proposition;
searching for a first word, from said text, in said lexicon;
analyzing the morphology of said word by identifying if a prefix or suffix has been added to a root word to form said first word;
identifying said propositions formed, in part, by said first word;
comparing said identified propositions and selecting propositions having at least one other common word forming at least a portion of at least two said propositions;
searching for a second word of said text in said lexicon;
analyzing the morphology of said word by identifying if a prefix or suffix has been added to a root word to form said second word;
identifying said propositions formed, in part, by said second word;
comparing said identified propositions and selecting propositions having at least one other common word forming at least a portion of at least two of said propositions;
comparing said selected propositions having said first word with said selected propositions having said second word; and
applying a fitness algorithm to determine the most applicable propositions of those compared. - View Dependent Claims (70, 71, 72, 73, 74, 75)
-
-
76. A ready access storage area for storing and processing proposition information in digital, analog, and other machine-readable formats, comprising;
-
at least one context structure comprised of a multi-dimensional network;
at least one structure comprised of a two-dimensional matrix; and
at least one structure comprised of a one-dimensional list of information. - View Dependent Claims (77, 78, 79)
-
-
80. A method for permitting users to enter language input, comprised of at least one language object, into a language processing system, comprising the steps of:
-
providing an input mechanism and an output mechanism;
providing at least one input object;
providing a permanent object storage structure having a plurality of objects therein and having information describing the locations of objects having a logic relationship to one another;
matching objects in said permanent object storage structure with objects presented as input objects by individually comparing each of said input objects with an object in permanent storage based on a predetermined similarity factor; and
returning information describing the location of all propositions in permanent storage relating to the matched input object to said output device. - View Dependent Claims (81, 82, 83, 84, 85)
-
-
86. A method for examining the objects of a knowledge molecule having x, r, y, and c objects and a weight, comprising the steps of:
-
providing a proposition having x, r, y, and c objects and a weight; and
providing a ready access storage area constructed and arranged to have at least one list for the storage of a plurality of objects and at least one weight; and
placing said objects of said proposition in said list in said ready access storage area with said weight. - View Dependent Claims (87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97)
-
-
98. A method for changing the weight of an object in a multi-dimensional context network in ready access storage, based on a match between one of the objects of a retrieved knowledge molecule and objects in a list in ready access storage, the method comprising:
-
providing a multi-dimensional network having at least one object stored therein, each said object having a weight assigned thereto, and said network also having at least one knowledge molecule stored therein, each said molecule having a weight assigned thereto;
retrieving said knowledge molecule; and
providing a formula for modifying a weight of said object, said formula for modifying weight of the object governed by said weight of said retrieved knowledge molecule. - View Dependent Claims (99, 100, 101, 102, 103, 104, 105, 106, 107)
-
-
108. A method for preparing an interpreter device to receive natural language input, comprising the steps of:
-
providing an interpreter device having a working storage area, a ready access storage area and a knowledge network, said ready access storage area having a plurality of objects each having at least one adjustable weight and vector in a context;
retrieving from said working storage area user defaults, parameters, predefined list and context information, words and propositions;
establishing in said ready access storage area reserved storage locations for said predefined lists and contexts;
storing in said predefined lists and contexts in said ready access storage area said predefined words and propositions retrieved from said working storage area; and
initiating simulated electrical potential flow in the knowledge network to adjust the weights of vectors of objects in said contexts in said ready access storage area.
-
-
109. A method for preserving context information while an interpreter device is not in operation, comprising the steps of:
-
providing an interpreter device having a ready access storage area, with context information therein, and a working storage area;
initiating a process, after receiving, but prior to completing a request to shut down an interpreter device, that keeps the device active until the process is complete;
transferring context information from said ready access storage area, to said working storage area for storage in said working storage area;
assigning a calendar date and time to said context information;
storing the calendar date and time applicable to said context information in working storage area with said context information; and
completing said shut down procedure.
-
-
110. A method for simulating human consciousness upon activation of an interpreter device, comprising the steps of:
-
providing an interpreter device having a working storage area, with context information stored therein, and a ready access storage area;
retrieving from said working storage area context information stored in said working storage area prior to shut down; and
storing said context information in said ready access storage area.
-
-
111. An interpreter device adapted to automatically obtain location information from a global positioning system constructed and arranged to generate information, comprising;
an interpreter device, having a ready access storage area, said interpreter device being constructed and arranged to receive information from a global positioning system. - View Dependent Claims (112, 113, 114)
-
115. An interpreter device adapted to automatically obtain date and time information from a time keeping device constructed and arranged to generate information, comprising;
an interpreter device, having a ready access storage area, said interpreter device being constructed and arranged to receive information from a time keeping device. - View Dependent Claims (116)
-
117. A device for interpreting language, comprising;
-
an interactive voice response system, said voice response system constructed and arranged to elicit information from a user and respond to a user;
a speech to data interpreter constructed and arranged to translate spoken information into data; and
a multi-dimensional interpreter constructed and arranged to analyze spoken information by analyzing the linguistic features, context, and relationships of the information and generate spoken responses and clarifying questions.
-
-
118. A method for interpreting language, comprising the steps of;
-
providing an interactive voice response system, said voice response system constructed and arranged to elicit information from a user and respond to a user;
providing a speech to data interpreter constructed and arranged to translate spoken information into data objects;
providing a multi-dimensional interpreter constructed and arranged to analyze said data objects by analyzing the linguistic feature, context, and relationships of the data objects;
providing a processor for further processing;
receiving input from a user with said interactive voice response system;
translating said input into data objects with said speech to data interpreter;
interpreting the linguistic features, context, and relationships of said data objects with said multi-dimensional interpreter to identify resulting information for further processing; and
sending the resulting information of the interpretation to a processor for further processing. - View Dependent Claims (119, 120, 121, 122, 123, 124)
-
-
125. A storage architecture for storing and processing logically organized information in any machine-readable format, comprising a multi-dimensional context network wherein the network has a plurality of dimensions arranged in a logically ordered scheme and each dimension is directly associated with adjacent dimensions by logical associations, comprising:
-
a first dimension used to store information representing the most general level of information stored in said multi-dimensional context network;
a second dimension used to store information related to information in said first dimension by a specific association representing a more specific level of information stored in said multi-dimensional context network than said first level;
a tertiary dimension arranged adjacent said second dimension, said tertiary dimension being outside said logically ordered scheme, used to store information related to information in said adjacent second dimension by a specific association representing the same level of specificity of information as in said second dimension;
any number of additional logically ordered dimensions, each of which is used to store information related to information in the dimension immediately above, in said ordered scheme, by a specific association representing a more specific level of information than that within the dimensions above;
at least one additional tertiary dimension adjacent to a logically ordered dimension, said tertiary dimension being outside said logically ordered scheme, each tertiary dimension being used to store information related to information in the adjacent hierarchical dimension by a specific association representing the same level of specificity of information but not part of the hierarchy stored in said multi-dimensional context network.
-
-
126. A method of matching each object component of a knowledge molecule with objects in a list in ready access storage and initiating further processing based on each match.
-
127. A knowledge appliance device with a human language interface designed to interpret human language speech and provide computing and communication services based on commands spoken by a user in a plain human language and interpreted by said device, the device comprising:
at least two interconnected data and instruction transfer busses comprised of at least one primary processing bus, one input bus, and one output bus. - View Dependent Claims (128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155)
-
156. A method of positively identifying a user, the steps of the method comprising:
-
placing a predetermined part of a user'"'"'s body in contact with a biometric reader having a scanner;
scanning a body part with the biometric reader and transmitting a digital biometric map of the scanned body part to an input/output subsystem;
transmitting said received digital biometric map from said input/output subsystem to a main processor;
storing the received digital biometric map in a shared memory area of said main processor;
invoking a program within the main processor to retrieve a list of digital biometric maps from a user database stored in an internal storage area of the main processor;
invoking a program within the main processor to compare the received digital biometric map with a plurality of known digital maps in the database; and
storing a name associated with the known digital biometric map in a user identification credential in the shared memory in the main processor if a match is achieved between the received digital biometric map and one of said plurality of known digital maps.
-
-
157. A method of positively identifying a user, the steps of the method comprising;
-
providing a sequence of predefined English words spoken by a user as if spoken to a human being;
providing a device for receiving a stream of sound waves comprising said sequence of words;
providing a microphone attached to said device for receiving said sound waves;
providing an input/output subsystem for transmitting said sound waves to a digital signal processor;
transmitting said received stream of sound waves to the digital signal processor;
creating a sound print of the words spoken by the user with the digital sound processor;
transmitting the received sound print with the input/output subsystem to a main processor;
storing the received sound print in shared memory within the main processor;
invoking a program to retrieve the list of sound prints from the user database stored in the internal storage area within the main processor;
invoking a program to compare the received sound print with a plurality of known sound prints within the database; and
storing the name associated with a known sound print in the within the shared memory area if a match is achieved between the received sound print and a known sound print.
-
-
158. A method of providing computing and communication services based on commands spoken by a user in a plain human language, comprising the following steps:
-
providing a user speaking human language words as if speaking to a human being who understands the human language, thereby creating a stream of sound waves generated by the user;
receiving the stream of sound waves generated by the user through a microphone attached to a device receives;
transmitting said stream of sound waves to a speech-to-text analyzer;
convert said stream of sound waves into a stream of text based on a plurality of speech conversion algorithms;
transmitting said stream of text to the an interpreter;
analyzing analyzes the meaning of said stream of text, by attempting to resolve ambiguity in the words and phrases and infer missing information using a plurality of multi-dimensional interpreter algorithms within an interpreter;
proceeding sequentially through the text and submitting preliminary results to a queue for a device controller subsystem to feed back to said speech-to-text analyzer to establish expectations and to improve the efficiency and accuracy of said speech-to-text analyzer;
generating human language text phrases of pertinent questions intended to complete or improve the interpretation of said stream of text, when said interpreter fails to achieve a sufficient interpretation of said stream of text;
submitting said pertinent questions phrased as text to the text-to-speech system to convert to said text to a spoken language question;
transmitting said spoken language question the speaker to generate output audible to the user;
providing an answer from the user by speaking into said microphone;
searching a list of functions for a function associated with said interpreted text, when said interpreter achieves a sufficient interpretation of said stream of text;
identifying a function matching said interpreted text;
executing program code associated with the matching function, when a matching function is found in said list of functions.
-
-
159. A device for processing text by an interpreter comprising:
-
a currently available computational device; and
an expansion module connected to said computational device, said expansion module used as an accelerator, said expansion module comprising a processor optimally configured to perform processes.
-
-
160. A method for expanding and improving the ability of an interpreter to correctly interpret input text, comprising the steps of:
-
adding a new object to a list of objects in a knowledge network;
adding a new molecular proposition to a network of propositions imitating the human process of learning a new fact;
changing at least one of the objects forming a molecular proposition imitating the human process of learning a new association between objects; and
adjusting a weight associated with a molecular proposition in the network of propositions imitating the human process of learning new information about a known fact. - View Dependent Claims (161, 162, 163)
-
Specification