JP2017508224A - System and method for recognition-based processing of knowledge - Google Patents

Abstract

A system for knowledge representation and application development includes a processor, a memory unit coupled to the processor, and a plurality of information units stored in the memory unit. An information unit is between an information unit and another information unit that defines one of the actions that the information unit performs when the information unit is impacted, ie activated. And at least one path describing a relationship that exists. The system impacts or activates the operator of the information unit, and further impacts, i.e., activates, another information unit to which the impacted information unit is connected via the path. A dynamic engine module executable by the processor is further provided for streaming. The information unit and path embody a knowledge representation schema, an information unit model, whereby an instantiation or instance of the information unit model is at least one of a domain, a set of observations, actual knowledge, and data. Refers to a specific set and assembly of model elements to accommodate one.

Description

  The field of the invention relates generally to automated processing of knowledge, and more specifically to applications of abstract models of information that provide certain advantages in access, visualization, discovery and processing.

  Information processing automation has been the goal of invention and mechanical design since humans conceived and built the oldest devices. Exploring advanced processing schemes is an area of great interest, has great potential for promoting scientific understanding, and continues to be adopted to generate wealth and benefit humanity.

  With the advent of computers, advances in available technologies such as computing power, process accessibility, disk storage and data networks have created interesting situations. These new technologies make it possible to implement many models of information processing that could only have been conceptual or thought experiments in the past. However, the same technology has also enabled the creation and storage of vast amounts of stored data and information, which in turn has created a need for advanced and automated techniques for processing large amounts of stored data.

  Information and cognitive modeling is beneficial in many scientific fields, including computer understanding, cognitive science, neurobiology, macro-conscious organization, education, decision support systems and basic physics. Modeling and automation of information processing, following specialized models, has been proposed as a key to understanding recognition and giving computers higher-order cognitive functions and understanding, perhaps discovered by the computer And make it possible to reach functions that are comparable to or exceeding human limits.

  The term Artificial Intelligence (AI) is intended to allow machines to exhibit human-like cognitive abilities, including thinking, understanding, reasoning, learning, problem solving, abstraction, observation and explanation. It covers a wide range of research and implementation. Any attempt to make a machine behave like a human or cause a machine to show intelligent behavior is often described as artificial intelligence.

  Many conventional approaches to artificial intelligence divide AI implementation into two areas: a knowledge base and a knowledge engine. A knowledge base captures knowledge about a particular domain in a machine-usable form. The knowledge engine applies a knowledge base. In other words, a knowledge base is similar to logic or rules, whereas a knowledge engine is functional instructions, algorithms and processing instructions that apply logic. However, many AI implementations obscure the distinction between logic and applications. For example, the logic may include processing instructions.

Designing and building a knowledge base is a major activity in knowledge engineering. The term knowledge engineering was first used in 1983 by Edward Feigenbaum. Feigenbaum defined knowledge engineering as an engineering discipline involved in integrating knowledge into computer systems to solve complex problems that normally require advanced human expertise. For the most part, domain experts provide knowledge for a knowledge base that can be converted to a knowledge base automatically or by human action. A particular problem is to encode knowledge in a format that makes it easily accessible to both the AI application and the maintainer and creator.

  One conventional approach for designing a knowledge base captures knowledge as a sequence of rules where each rule comprises a predecessor and a successor. The predecessor is a series of conditions that, when satisfied, trigger an action specified in the subsequent part. The rules are then processed against the set of input data to obtain one or more final conclusions. Although rule-based expert systems have achieved success in a limited domain, the challenge is to capture domain knowledge in a set of rules. Most conventional systems have required domain experts to work with a knowledge engineer to capture the rules. Another challenge facing rule-based systems is to apply rules in a convenient manner and limit the use of rule-based systems in time-critical systems. It is often debated whether human experts use the rules and correspondingly the automated application values of the rules in reaching the conclusion.

  The traditional approach to building a domain knowledge model is training in which an automated observer examines a series of events and conditions and, as a result, builds a domain knowledge model based on the observations. In supervised learning (also called reinforcement training), a teacher or other source of expertise guides the training by providing accurate results. In unsupervised learning, an automated learning system looks for patterns or features in the dataset without guidance. Some conventional systems have attempted to automatically construct rules from observed data.

  Another approach to artificial intelligence employs artificial neural networks. A network of biological neurons makes up the majority of the brains of humans and other organisms. Artificial neural networks use similar principles, using computing algorithms or hardware to implement interconnecting artificial nodes or neurons. Each node can receive a stimulus from its input network and, when properly stimulated, it “fires” and sends the stimulus to the node connected to its output. Connection and excitement thresholds determine network functionality. Neural networks are sometimes generalized in the field of connectionism.

  Early work in neural networks focused on emulating biological system structure and function, but most modern work is based on signal processing and statistics. The problem with neural networks is the requirement that the neural network be trained on the knowledge domain. Training can be time consuming and the knowledge gained is specific to the training data. Thus, artificial neural networks do not generalize or gain the ability to deal with data outside the training domain. Neural networks have been successful in limited domains such as control systems.

  A classifier is an automated system for classifying a set of data, recognizing patterns in the data, and generally automatically identifying relationships in the data set. Such relationships can include interrelationships, rankings, clusters and features. Many classifiers are implemented using kernel methods or support vector machines. The kernel method has had some success in areas such as handwriting appraisal and automatic character recognition.

Verification and validation is another challenge for computer-based intelligent systems. Many critical systems require intensive testing to verify that the system meets specifications and to validate that the system works as expected before it is deployed. This test, which is common in automated recognition implementations, is difficult when knowledge is captured in a form that is not easily understood by humans and the behavior of the system is complex or non-linear.

Other relevant areas are cognitive modeling, information theory, and Kolmogorov complexity.
The present invention relates to a computer-implemented model that describes the elements, structures, and processes that exist during intelligent thinking and information processing.

  The model and its computer implementation are intended to be based on its simplicity (consisting of only a few elements) and universality (the ability to accurately represent the very diverse experiences faced in this area) Has been. The methodology has both a formal description and application to several areas of human effort, including science and information processing technology.

  The foregoing features and objects of the invention, as well as other features and objects, and techniques for obtaining them, will be more fully understood by reference to the following description of embodiments of the invention, considered in conjunction with the accompanying drawings. It will become apparent and the invention itself will be better understood.

1 is a schematic diagram of a computer system that implements an embodiment of the present invention. FIG. 3 is a dendrogram illustrating class configuration and related components, according to one embodiment of the present invention. FIG. 4 is a dendrogram illustrating the configuration of elements and related components according to one embodiment of the present invention. FIG. 3 is a tree diagram illustrating path configuration and related components, in accordance with one embodiment of the present invention. FIG. 3 illustrates an embodiment of the present invention applied to a domain of kinematics of point objects in a closed system. FIG. 3 illustrates an embodiment of the present invention applied to the kinematics domain of point objects in a closed system. FIG. 3 illustrates an embodiment of the present invention applied to the kinematics domain of point objects in a closed system. FIG. 3 illustrates an embodiment of the present invention applied to the kinematic domain of point objects in a closed system. FIG. 3 illustrates an embodiment of the present invention applied to the kinematics domain of point objects in a closed system. FIG. 3 illustrates an embodiment of the present invention applied to the kinematics domain of point objects in a closed system. FIG. 3 illustrates an embodiment of the present invention applied to the kinematics domain of point objects in a closed system. The figure which illustrates one Embodiment of this invention applied with respect to the domain of business management. The figure which illustrates one Embodiment of this invention applied with respect to the domain of business management. The figure which illustrates one Embodiment of this invention applied with respect to the domain of business management. The figure which illustrates one Embodiment of this invention applied with respect to the domain of business management. The figure which illustrates one Embodiment of this invention applied with respect to the domain of business management. The figure which illustrates one Embodiment of this invention applied with respect to the domain of business management. The figure which illustrates one Embodiment of this invention applied with respect to the domain of business management. FIG. 3 illustrates an embodiment of the present invention applied to the domain of object identification in an array of single valued pixels. FIG. 3 illustrates an embodiment of the present invention applied to the domain of object identification in an array of single valued pixels. FIG. 3 illustrates an embodiment of the present invention applied to the domain of object identification in an array of single valued pixels. FIG. 3 illustrates an embodiment of the present invention applied to the domain of object identification in an array of single valued pixels. FIG. 3 illustrates an embodiment of the present invention applied to the domain of object identification in an array of single valued pixels. FIG. 3 illustrates an embodiment of the present invention applied to the domain of object identification in an array of single valued pixels. 6 is a flow chart illustrating impact processing within a single unit, according to one embodiment of the invention. 5 is a flow chart describing one embodiment of a process step for conversion to a unit model representation. The flowchart explaining one Embodiment of the process of the problem-solving method of this invention. 2 is a flow chart illustrating one embodiment of the present invention related to discovery. 6 is a flow chart illustrating a mapping process according to one embodiment of the present invention. The flowchart explaining the payment list creation process which concerns on another embodiment of this invention. 1 is a schematic diagram of a network system in which an embodiment of the present invention can be utilized. For optional input devices (eg, keyboard, mouse, touch screen, etc.) and output devices, hardware, network connections, one or more processors, and data and modules that can be utilized in conjunction with embodiments of the present invention FIG. 2 is a block diagram of a computing system (such as a server or client, or both, as appropriate) having memory / storage, etc.

  Corresponding reference characters indicate corresponding parts throughout the several views. Although the drawings represent embodiments of the present invention, the drawings are not necessarily drawn to scale and some features may be exaggerated to better illustrate and explain the present invention. . The illustrations set forth herein illustrate one embodiment of the invention in one form, and such illustration should not be construed as limiting the scope of the invention in any way. Absent.

  In the description and drawings that follow, a few components and connections may be shown for ease of explanation and illustration. The number of components used herein is for example purposes only. It should be understood that these examples are not intended to illustrate the ultimate functionality of the present invention, including the number of possible components, the number of instances or interconnections. The embodiments disclosed below are not intended to be exhaustive or to limit the invention to the precise forms disclosed in the following detailed description. Rather, the embodiments have been chosen and described so that others skilled in the art can utilize the teachings.

The detailed description that follows is presented in part in terms of algorithms and symbolic representations of operations (operations) on data bits in computer memory that represent alphanumeric characters or other information. A computer typically includes a processor for executing instructions and a memory for storing instructions and data. If a general purpose computer has a sequence of machine-encoded instructions stored in its memory, the computer operating on such an encoded instruction may be a specific type of machine, i.e., by a sequence of instructions. It can be a computer that is specifically configured to perform the operations embodied. Some of the instructions can be adapted to generate signals that control the operation of other machines, and thus operate through those control signals to convert material far from the computer itself. These descriptions and representations are the means used by those skilled in the data processing arts to most effectively convey the substance of their work to others skilled in the art.

  The algorithm is here also generally considered to be a self-consistent sequence of steps leading to the desired result. Such a process is a process that requires physical manipulation of physical quantities. Usually, though not necessarily, such quantities are stored, transmitted, converted, combined, compared, otherwise compared to electrical pulses or electrical signals or magnetic pulses or magnetic signals that can be manipulated. Take the form. Primarily for general usage reasons, such signals are represented by bits, values, symbols, characters, display data as a reference to the physical goods or physical expression in which such signals are embodied or represented. It has proven convenient at times to refer to as terms, numbers, etc. It should be noted, however, that all of these and similar terms are to be associated with the appropriate physical quantities and are only used here as convenient labels that apply to such quantities. Should.

  Some algorithms may use data structures for both inputting information and generating desired results. Data structures greatly facilitate data management by data processing systems and are not accessible except through sophisticated software systems. The data structure is not the information content of the memory, but rather the data structure represents specific electrical structural elements that convey or manifest the physical organization of the information stored in the memory. More than just abstraction, data structures accurately represent the physical characteristics of complex data, often data modeling of related items, while providing improved efficiency in computer operations. Specific electrical or magnetic structural elements.

  Furthermore, the operations performed are often referred to in terms such as comparison or addition that are commonly associated with mental operations performed by a human operator. Such a function of a human operator is not necessary or desirable in most cases in any of the operations described herein that form part of the present invention. The operation is a machine operation. Useful machines for performing the operations of the present invention include general purpose digital computers or other similar devices. In all cases, a distinction should be recognized between the method operation in operating a computer and the method of operation itself. The present invention is a method and apparatus for operating a computer in processing electrophysical signals or other (eg, mechanical, chemical) physical signals to produce other desired physical expressions or signals. About. The computer runs on a software module, which is a signal stored on a medium that represents a series of machine instructions that allow the computer processor to execute machine instructions that implement the algorithm steps. It is a set. Such machine instructions may be actual computer code that the processor interprets to implement the instructions, or alternatively, higher order instructions that are interpreted to obtain the actual computer code. It can be an instruction coding. A software module may also include hardware components, in which case some aspects of the algorithm are performed by the circuit itself rather than as a result of an instruction.

The present invention also relates to an apparatus for performing such operations. The apparatus may be specifically constructed for the required purpose, or may comprise a general purpose computer selectively activated or reconfigured by a computer program stored in the computer. . The algorithms presented herein are not inherently related to any particular computer or other apparatus unless explicitly indicated as requiring specific hardware. In some cases, a computer program may require other hardware or programming to interact with other programs or equipment through signals configured for a particular protocol that may or may not require. Can communicate with or be associated with. In particular, various general purpose machines may be used with programs written in accordance with the teachings herein, or to construct more specialized equipment to perform the required method steps. May prove to be more convenient. The required structure for a variety of these machines will appear from the description below.

  In the following description, some frequently used terms have special meaning in this context. The terms “window environment”, “running in a window”, and “graphical user interface operating system” are used on a video display, such as in a bounded area on a raster-scanned video display. Information is manipulated and used to represent the computer user interface that is displayed. The terms “network”, “local area network”, “LAN”, “wide area network” or “WAN” mean two or more computers connected so that a message can be sent between the computers. . In such a computer network, typically one or more computers are used to operate a “server”, a mass storage device such as a hard disk drive, and a peripheral device such as a printer or modem. It operates as a computer having communication hardware. Other computers, referred to as “workstations”, provide a user interface so that users of a computer network can access network resources such as shared data files, common peripheral devices, and inter-workstation communications To do. A user activates a computer program or network resource to create a process that includes both the general behavior of the computer program, as well as the specific behavioral characteristics determined by the input variables and its environment. Similar to a process is an agent (sometimes called an intelligent agent) that collects information or performs some other service on some regular schedule without user intervention It is. Typically, the agent uses parameters typically provided by the user to search for locations at the host machine or at some other point on the network, collect information about the agent's purpose, and Present information periodically to the user.

  The term “window” and the associated terms such as “window environment” or “running in a window” as defined above are several of those available from Microsoft Corporation, Redmond, Washington. Refers to a computer user interface exemplified by a window system. Other window computer interfaces are available, for example, from Apple Computers Incorporated of Cupertino, CA and as a component of the LINUX® operating environment. In particular, it should be understood that the use of such terms in the description herein does not imply limitations on any particular computing environment or operating system.

The term “desktop” refers to a particular user interface that presents a menu or object display with associated settings for a user associated with the desktop. When the desktop accesses network resources, this typically requires that the application program be run on a remote server, and the desktop calls the application program interface or API. , Allowing the user to provide commands to network resources and observe any output.

  The term “browser” is not necessarily obvious to the user, but sends messages between the PDF and the network server and displays to the network user and interacts with the network user. A program that is responsible for Browsers are designed to take advantage of communication protocols for the transmission of text and graphic information over a worldwide network of computers, namely the “World Wide Web” or simply the “Web”. Examples of browsers compatible with the present invention are created by Internet Explorer sold by Microsoft Corporation (Internet Explorer is a trademark of Microsoft Corporation), Opera Software ASA (Opera Software ASA). Opera browser programs, or Firefox browser programs distributed by Mozilla Foundation (Firefox is a registered trademark of Mozilla Foundation). Although the following description describes in detail such operations from the perspective of a browser graphic user interface, the present invention provides a text-based interface having many of the features of a graphic-based browser, Alternatively, it can be implemented using a voice activated interface or even a visually activated interface.

  The browser displays information that is formatted in a standard generic markup language (“SGML”) or hypertext markup language (“HTML”), both of which are special ASCII text code A scripting language that embeds non-visual code in text documents through use. Files in these formats can be easily transmitted over computer networks including global information networks such as the Internet, allowing browsers to display text, images, and play audio and video recordings. The web utilizes these data file formats in conjunction with web communications protocols for transmitting such information between the server and the workstation. The browser can also be programmed to display information provided in an extensible markup language ("XML: extensible Markup Language") file, which contains several document type definitions ("DTD: Document Type Definitions"). "), And is therefore inherently more common than SMGL or HTML. An XML file can be similar to an object because the data and stylesheet format are included separately (the format may be thought of as a way to display information, so the XML file And associated methods).

  The term “personal digital assistant” or “PDA” as defined above refers to any handheld portable device with computing, telephone, fax, e-mail and network functions. The term “wireless wide area network” or “WWAN” means a wireless network that serves as a medium for transmission of data between a handheld device and a computer. The term “synchronization” refers to the exchange of information between a handheld device and a desktop computer via wire or wirelessly. Synchronization ensures that the data on both the handheld device and the desktop computer is the same.

  In wireless wide area networks, communication occurs primarily through the transmission of wireless signals over analog networks, digital cellular networks, or personal communication services (“PCS”) networks. The signal may be transmitted through microwaves and other electromagnetic waves. The electromagnetic waves used for communication may include “light” waves at or near visual frequencies transmitted through free space or using “optical fibers” as waveguides. Currently, most wireless data communications include code division multiple access (“CDMA”), time division multiple access (“TDMA”), global system for mobile communications (“GSM”), Cellular digital packet data (2) used on cellular systems using second generation technologies such as personal digital cellular ("PDC") or on advanced mobile phone services ("AMPS") Occurs through packet data technology on analog systems, such as "CDPD"). The term “wireless application protocol” or “WAP” refers to a universal specification for facilitating the delivery and presentation of web-based data using a small user interface on handheld and mobile devices. .

  The term “real time” (also referred to as “real time”) or “almost real time” refers to a system design approach that uses timing as the primary design objective. Specifically, the real time system completes one or more operations within a time interval that satisfies a predetermined criterion. The term may also be used to refer to the action being performed, eg, “real time update”. The time interval criteria may be a specific amount of time or defined as opposed to another non-real time system, sometimes referred to as a “batch” system or an “offline” system. Also good. It can be seen that the time interval is determined by requirements that vary from system to system. For example, a real-time aircraft real-time control system may need to respond in microseconds, whereas a real-time water level adjuster is acceptable to be updated every few hours. It can be. Systems that provide a “real-time response” in human user interaction are fast enough to allow interactive or “live” use of the system without an unpleasant delay, Means that the user receives a response to the input.

  In this description, real-time transaction processing is designed to quickly complete operations that affect system data, and the resulting modified data is stored in an offline synchronization process. Is made available to other system components as quickly as possible. The exact timing of such a system depends on many factors, such as processing time and the propagation of data over the network, but the salient feature is the rapid availability of data that has changed as a result of a transaction event It can be seen that it is sex.

  FIG. 12A is a high-level block diagram of a computing environment 1200 according to one embodiment. FIG. 12A illustrates a server 1210 and three clients 1212 connected by a network 1214. For simplicity and clarity of illustration, only three clients 1212 are shown in FIG. 12A. Embodiments of the computing environment 100 may have thousands or millions of clients 1212 connected to a network 1214, eg, the Internet. A user (not shown) can send a message through server 1210 and its associated communication equipment and software (not shown) through network 1214 and receive a message from client 1212. Software 1216 may be operated on one of them.

FIG. 12B illustrates a block diagram of a computer system 1211 suitable for implementing a server 1210 or client 1212. The computer system 1211 includes a central processing unit 1215, a system memory 1217 (typically RAM, but the system memory may include ROM, flash RAM, etc.), an input / output controller 1218, and an audio output interface 1222. An external audio device such as a speaker system 1220 via, an external device such as a display screen 1224 via a display adapter 1226, a serial port 1228 and 1230, a keyboard 1232 (interfaced using the keyboard controller 1233), Storage interface 1234, disk drive 1237 operating to accept floppy disk 1238, fiber channel network A host bus adapter (HBA) interface card 1235A that operates to connect to 1290, a host bus adapter (HBA) interface card 1235B that operates to connect to SCSI bus 1239, and an optical disk 1242 It includes a bus 1213 that interconnects the major subsystems of the computer system 1211, such as an optical disk drive 1240 that operates on the computer. Mouse 1246 (or other point-and-click device coupled to bus 1219 via serial port 1228), modem 1247 (coupled to bus 1212 via serial port 1230), and network An interface 1248 (coupled directly to bus 1213) is also included.

  Bus 1213 allows data communication between central processing unit 1214 and system memory 1217. The system memory 1217 may include read only memory (ROM) or flash memory (both not shown) and random access memory (RAM) (not shown) as described above. RAM is typically main memory into which an operating system and application programs are loaded. ROM or flash memory may include a basic input / output system (BIOS), among other software code. The BIOS controls basic hardware operations such as interaction with peripheral components. Applications that reside on computer system 1211 typically include a hard disk drive (eg, fixed disk 1244), an optical drive (eg, optical drive 1240), a floppy disk unit 1237, or It is stored on a computer readable medium such as another storage medium and accessed via a computer readable medium. The application may also be in the form of electrical signals that are modulated according to the application and data communication technology when accessed through the network modem 1247 or interface 1248 or other telecommunications equipment (not shown). .

  The storage interface 1234 may be connected to a standard computer readable medium for storage and / or retrieval of information, such as a fixed disk drive 1244, similar to other storage interfaces of the computer system 1210. . Fixed disk drive 1244 may be part of computer system 1211 or may be separate and accessed through other interface systems. Modem 1247 may provide a direct connection to a remote server via a telephone line or the Internet via an Internet service provider (ISP) (not shown). The network interface 1248 may provide a direct connection to a remote server via a direct network link to the Internet via a point of presence (POP). Network interface 1248 may provide such connections using wireless techniques including digital cellular telephone connections, cellular digital packet data (CDPD) connections, digital satellite data connections, and the like.

Many other devices or subsystems (not shown) can be connected in a similar manner (eg, document scanners, digital cameras, etc.). Conversely, not all of the devices shown in FIG. 12B need be present to implement the present disclosure. Devices and subsystems may be interconnected differently than the approach shown in FIG. 12B. The operation of a computer system such as the computer system shown in FIG. 12B is readily known in the art and will not be discussed in detail in this application. Software sources and / or object code for implementing the present disclosure include a computer, such as one or more of system memory 1217, fixed disk 1244, optical disk 1242, or floppy disk 1238. It can be stored in a readable storage medium. The operating system provided on the computer system 1211 is MS-DOS (registered trademark) (MS-DOS is a registered trademark of Microsoft Corporation, Redmond, Washington), WINDOWS (registered trademark) (WINDOWS (registered trademark)). Are trademarks of Microsoft Corporation of Redmond, Washington), OS / 2 (registered trademark) (OS / 2 is International Business Machines Corporation of Armonk, NY). ), UNIX (registered trademark) is a trademark of X / Open Company Limited of Reading UK. (Registered trademark of X / Open Company Limited), Linux (registered trademark) is a registered trademark of Linus Torvalds, Portland, Oregon, or other It can be a type or version of any known or developed operating system.

  Further, with respect to the signals described herein, one of ordinary skill in the art will be able to transmit the signal directly from the first block to the second block, or the signal may be changed between blocks (eg, amplified, (Attenuated, delayed, latched, buffered, inverted, filtered, or otherwise modified). While the signals of the above-described embodiments are characterized as being transmitted from one block to the next, other embodiments of the present disclosure may have signal information and / or functional aspects transmitted between blocks. As long as such a signal is directly transmitted, a modified signal may be included. To some extent, the second input in the second block is output from the first block due to physical limitations of the circuits involved (eg, some attenuation and delay necessarily exists). It can be conceptualized as a second signal derived from one signal. Therefore, in the present specification, the second signal derived from the first signal includes the first signal, or is caused by a circuit limitation, or information and / or final of the first signal. Any change to the first signal is included, whether due to the passage of other circuit elements that do not change the functional aspects.

  Many of the features and descriptions herein refer to an information model referred to as a “unit model”. The design and structure of the unit model is itself part of this disclosure, and its features are the basis for many of the other features and descriptions of embodiments disclosed herein. It is.

The description herein can be broadly divided into three aspects. First, the unit model is an abstract set of elements, methods, and relationships. Second, a unit model instantiation or instance refers to a specific collection and assembly of model elements that are collected in response to a domain, a set of observations, actual knowledge, data, and the like. A model instance can be said to be a model of a domain. Third, the implementation of a unit model refers to the mechanical implementation of a model or model instance, for example, as programs and data in a digital computer. It should be appreciated that references using these three aspects can be used interchangeably. For example, a reference to a model may also refer to an implementation of the model and method on a computer. As another example, a reference to an instance of a model can be viewed as a reference to a representation of the instance in computer memory.

  A unit model (also referred to herein as a “model” or “information unit model” or “knowledge unit model”) is a model consisting of an abstraction of structures and processes corresponding to intelligent information processing. It is. The model is based on a single atom called an “information unit” or “knowledge unit”. All areas of information can be described by an array of information units and the relationships between them. Some embodiments of the present invention include different configurations of such knowledge units.

  An information unit is a representation. An information unit can represent anything. An information unit may represent an actual object such as a car, or an information unit may represent an abstract thought such as an idea. An information unit may also represent a sequence of information units.

  2A-2C are tree diagrams illustrating the structure of the information unit and its components. Referring to FIG. 2A, an information unit 202 is shown. Unit 202 has several associated attributes including a unique ID 204, a name 206, an element list 208 and a list of paths 210.

  Every unit typically has a unique identifier. It is possible to have the same identifier, but this may imply the possibility of a secondary key, or incorrect unit call.

  Every unit may have a name or several names, which are text strings used to help identify the meaning of the unit. For example, a unit may represent a blue car with an ID of 123xyz, but the name may be “blue car”. A unit may represent a mathematical variable that represents the price of a bond. The ID can be 567 mno. This unit may have two names, the first name is “bond price” and the second name starts with “BP”, which is more suitable for display in the equation. The name of a unit is not necessarily unique and is only used to help human users identify what the unit represents.

  The element list 208 represents a sequence of elements. FIG. 2B illustrates the structure of the element. Element 230 is a structure that identifies another information unit. There are several ways in which element information units can be identified. The approach is 1) referenced directly by its id 232, 2) indirectly referenced as an information unit operator target 234, and 3) indirectly referenced as an information unit source target 236. Can be. Further, the information unit referenced by the element may be an element referenced by the information unit. Elements can be identified as symbol elements or symbol sequences for use in mapping operations. In addition, an element may refer to a unit by any of the above methods, for example, appended with a path that may point to an attribute of the class.

  The path list 210 is a list of paths that the unit has with other information units. The path structure is illustrated in FIG. 2C. Path 250 consists of source unit 252, relationship unit 256, destination unit 254, and resistor 258.

Truth value 212 is a value, such as a binary or real value, that indicates the condition or degree that the unit represents the reality or success of the operator of the unit.
The desirability factor 214 is a real value that indicates the degree to which the unit represents something desirable or undesirable for the unit model instantiation embodiment.

  The operator 216 indicates an action to be performed when a sufficiently large impact is received. The actions taken by the operator can create, modify, or delete information units, create, modify, or delete paths, and are specific to model embodiments, e.g., input devices or output devices You can call the actual operator that manipulates. An operator may more generally be any action that can act on a sequence of units, i.e., reorder units, and the presence of a particular unit or set of units in another sequence of units. Map and check.

  Many operators can be created in embodiments of the present invention. The listed operators are provided as examples of possible operators for units, atom values, and other objects in some embodiments of the invention. Such operators include, but are not limited to:

Createunit-creates a new unit. The new unit is the target of the unit.
Copy-Copy units, sets of units and their relationships (eg, classes) Copyonto-Copy source unit relationships and apply them to destination units Deleteunit-Delete units Shock-Shock unit Send to Fact-an operator that does nothing. The element only serves to represent a sequence and is therefore called Intercept? -Does one list have elements in common with another list? Intersec-ex? -Whether all elements of one list are elements of another list Loop (list, element)-cycle for each element of the list, with the first element of [element] relative to the current element Set. The next time the loop receives an impact, the loop sets [element] to the next element in the list's element list.

Reset (loop)-Resets the loop unit to set its element variable to the first element in the list the next time the loop is impacted.
Mappings—Applies a map to the elements of the target unit and, if successful, sets a virtual unit for one or more of the target elements. If successful, set the truth value to true, otherwise set it to false.

    Mappaths—provides a map for the path of the target unit and, if successful, sets a virtual unit for one or more of the target path elements. If successful, set the truth value to true, otherwise set it to false.

Addpaths: [Unit] [Source] [Relation] [Destination]-Add path to unit Deletepath-Delete path from unit Clear-Remove all elements from the list of units and clear the virtual unit.

Same Unit? : [Unit 1] [Unit 2]-Are Unit 1 and Unit 2 the same? : [Unit 1] [Unit 2]-Do Unit 1 and Unit 2 have similar units? Set [Unit 1] [Unit 2] [Unit 3] * [Unit 4]-[Unit 1] to [Unit 2] [Unit 3] and [Unit 3] Set the element of Unit4] Real: [Operator] [Parameter1. . n] —An operator that invokes the actual operator specified by [Operator]. The actual operator is anything that interacts with anything outside the unit model, for example a display, storage device, or input device.

  In addition to operators for units, atoms representing real numbers, input functions, processes, and other environmental aspects can be operated on by one or more of the following actual operators.

a = b + c, a = b-c, a = b * c, a = b / c-simple arithmetic operations Cos, Sin, Tan, ArcCos, ArcSin, ArcTan, Cosh, Sinh, Tanh-trigonometric operations Ln, Log, Power-Logarithm IsNumber? -Does the unit element represent a number>, <, <=,> =,! =, =? -Simple arithmetic comparison operator Div-quotient and remainder operator InputText-get text from the user Message-set the text of the message to be shown on the display MessageYesNo-ask the user to click the Yes or No button GetSelected-indicates a message box to be requested-asks the user to select a unit from the application user interface ApplicationMenu-adds an item that can access a specific process to the application main menu-UnitMenu-unit Intelligently add items that can access a specific process to the right-click menu of Get Omment - obtaining comments unit ShowUnit - showing the interaction foam having a control that represents a set of units - units Form shown on the display. This set is usually a class, class instance or list.

Import-Import a delimited text file CountElements-Count the number of elements in a unit FindParentProcess-Find the process of which a process is a member SetWorkingKB-Set the system's WorkingKB variable. WorkingKB is used for multiple purposes, in particular SetUnitKB used to set the domain of a new unit to be created-Set the domain of a specific unit SetTrueValue-Set the truth value of a unit Sleep An operator IsEmpty that causes an impact flow to pause for a specified number of milliseconds? -Does the unit have any elements? If a unit has an element, the operator loads the knowledge domain from the LoadKB-storage device that sets the truth value of the unit to false.

  A unit may have an operator target unit 215 that is subject to the operation of this unit. The unit 202 may also have a source target unit 213 that identifies the unit that activates the unit 202.

  The unit model has a small set of components and is easy to work with. An exemplary representation of the unit model is provided in FIGS. 2A-2C. The model is also very general. The model consists of a single general structure (information unit) and is sufficient to model any set of events generated by the source. Models are also common in their ability to accurately represent the very diverse experiences they face. An unlimited number of event combinations can be represented using a model.

  The unit model may be employed as a source observer, where the source creates a series of “events”. Each observed event is then instantiated as an information unit in the model. Events are typically represented in the model nomenclature by the two-character sequence “Ev”. An event is an information unit in the model.

  If multiple events are observed and multiple corresponding information units are instantiated in the model, the “Relation” between units can also be instantiated. Examples of such relationships can address the timing between events, where the relationship can be “before”, “after”, or “simultaneous”. Another example of a relationship between units is “different” indicating that the events are not identical. The relationship is typically represented by a lowercase “r” in model nomenclature. Information units can also be used to represent relationships between units in a model.

Two units, together with the relationship between the two units, are defined as “paths”. A path can also be represented by an information unit in the model.
A “state” is a set of instantiated events. Thus, each observed event creates a new state in the model. A state is typically represented by a lowercase “s” in model nomenclature. The state is also an information unit in the model.

  A “situation” is defined as a continuous set of states. Specifically, a situation is a sequence of states of related information units. A situation is typically represented by an uppercase “S” in model nomenclature. The situation is also an information unit in the model.

  If one particular state is always followed by another particular state, the two states are said to have a “conditional relationship”. The conditional relationship is typically expressed by two character strings “->” in the model nomenclature. Conditional relationships are also information units in the model.

  A “symbol unit” is a unit that can take on the value of any other unit. Symbolic units are useful for generalization when more than one event matches a case. A symbolic unit is typically represented in a model nomenclature by a string with an underscore (“_”) followed by an integer, and is used in operations involving mapping. A symbol unit may be constrained to represent a single unit, or may represent a sequence of units.

Note that each conditional relationship has a unit on which the conditional relationship depends (preceding condition) and other units that are the result of satisfying this preceding condition. Thus, the resulting unit does not represent only the sequence of units, but also shows how the next state differs from the previous state. The result is therefore also the subject of change.

  In some embodiments, the result includes the creation of a new unit. In other embodiments, the result instead modifies the existing unit. Thus, satisfaction of a conditional relationship may create, modify, or remove an information unit (target unit) as defined by its operator.

  Each information unit may have a relationship to one or more other information units, which may possibly include a sequence stored as a path. The relationship between an information unit and another information unit or target sequence can be equivalence, mapping or abstraction.

  If the information unit includes a plurality of units, the element sequence may be ordered according to various criteria in various embodiments, or the sequence may not have a specific order. For example, the sequences may be ordered or sorted according to the creation date or similar temporal criteria. Therefore, every experience can be expressed as a set of information units.

A “library” is a collection of domains.
A “common flow” is a flow that is free and unrestrained. A “bipolar flow” is a flow from two displaced information units.

Encapsulation is a process in which multiple units are referred to as a single unit. Examples of encapsulation are lists, processes, and domains.
A persistent event or a set of persistent related events is called an “object”. An object may be a list, a class instance, and may correspond to an actual physical thing. A “list” is defined as an object composed of a sequence of elements.

A “process” is an encapsulation of a set of units that have a conditional relationship (usually linear) to each other. The process directs one or more impacts to the sequence of units.
A “domain” is a domain of knowledge. A synonym for a domain is “universe”. A domain may include objects, classes, information units, and other unit model elements.

  A set of objects that all have a similar set of relationships is called a “class”. A class is a virtual set of units that represents a set of information units where all of the represented information units have similar characteristics. There is no limit on the number of information units in a class. Similarly, in this context, it means that the set of relationships or the set of features is the same at some level of abstraction.

A member of a class is called an “instance” of the class. A similar relationship is referred to as an attribute relationship, and an event forming the relationship is referred to as a class “attribute”.
Every class can have a set of processes that encapsulate all the conditional relationships that can exist between class attributes. Such a process is called the “principle” of the class. In general, a class may have any number of attributes and principles.

The attribute can be a text string, a text string constrained by an optional list, the class itself or a list. For example, the name attribute is a text string type. The gender (attribute) of a person (class) is an example of a text string attribute constrained by a list of “male” or “female”. Employee (list attribute) is an example of a list attribute where each list item is a class employee, where employee is a company attribute.

  We have described objects, classes, lists and processes. However, information units can be arranged in many more types of configurations. Information units that are random sequences of other information units may be created. Because of the correlation with the observed state, some of the units created may represent a conditional relationship that can be considered as a predictor of future states. However, since past performance does not necessarily indicate future results, a unit representing a conditional relationship may or may not be associated with a true or desired truth value.

  Thus, each unit has a binary characteristic that indicates its association to observation, and this characteristic is referred to as its “truth value” or “TV”. The truth value may be “true” or “false”, or in alternative embodiments may be a real quantity. Truth values may also represent the success or value of unit operators in some examples. The truth value of a unit can be changed by the action of an operator of the same unit or an operator of a different unit. In other embodiments, the truth value may be a non-binary real value that represents a fuzzy logic evaluation of this property of the unit. Such a fuzzy value may be a real number, for example, an integer value in the range of 0 to 1, a range between a predetermined minimum value and a predetermined maximum value, or other set of possible values.

Each unit also has a desirability factor that can be an integer or real valued property corresponding to the desirability of that unit relative to other units.
As described, information units can be constructed and added to the universe as a result of observation. An additional property of the information unit is called an “operator”. Operators apply impacts, perform logical or set operations on information units as described below, order such information units, sequence them, and perform other processing functions , Giving the unit the ability to create and modify a set of information units of the unit itself. Thus, a unit may add, remove, copy, and edit its own set of information units, including adding, removing, copying, or changing individual parts of an information unit. The operator reflects the observed subject of change.

  Three additional change operators are applied to the information unit: impact operator, path resistance setting, and threshold setting. The impact operator gives an “impact” to the unit, thus allowing a selective start of the flow. The path resistance operator sets the “resistance” characteristic of the path. In one embodiment, the resistance is a positive actual quantity. In one embodiment, the resistance characteristic is similar to fluid resistance, mechanical resistance, or electrical resistance in that the resistance resists the flow of energy and wastes energy. Therefore, the resistance characteristic enables selective control of the flow by opening and closing a path through which energy flows. In one embodiment, the resistance characteristics of the path change over time. For example, the resistance of a unit may decrease as energy flows through the path. Thus, the path may include a history memory of experienced flows. The past history of the flow through the path can affect the resistance to the current flow. In other embodiments, resistance can be defined as a constant characteristic of a particular unit.

Each unit has a characteristic known as its “threshold”. The threshold quantifies the number and / or amount of impact required to activate the unit operator. In some embodiments, a single “impact” activates a unit, and the unit consequently impacts an probably unlimited number of “destination units”. In other embodiments, multiple impacts are required to activate the operator and initiate an impact along all of its conditional paths. In one embodiment, multiple impacts must arrive within a set period of time. In another embodiment, multiple impacts are accumulated until the cumulative amount of impact reaches or exceeds the unit activation threshold.

  One exemplary embodiment of a system that uses a unit model is shown as system 100 in FIG. Data bus 102 interconnects processor 108, input device 102, output device 104, communication interface 106, and memory 110. The memory 110 stores various items used by the processor 108 to implement the methods described herein, and includes an operating system 112, a unit model module 114, a model visualization module 116, and the like. A model editor module 118, a model observation module 120, an operation module 122, and a dynamics module 124, all of which have access to the unit model storage 126.

  In a network with many interconnected information units, the number of possible operators that can be activated is unlimited. If this happens, many new units are added to the total set. In accordance with some of the various embodiments of the present invention, applying the concept of finite energy and resistance limits the number of possible activations.

  Another important aspect of the unit model is named flow. This aspect of the model is called dynamics. A “flow” is defined as a trajectory of activated units and as a path connecting these units. The units included in the flow path are different from the units outside the flow. This is because each unit involved experiences some dynamic characteristic flow through that unit. This dynamic characteristic is called “energy”. Therefore, a unit that can be impacted with sufficient energy shall activate its operator. Any unit can be a precondition in a conditional relationship. The kinetics of applying energy explains how the result is activated. Conditional relationships operate as a path for transferring energy from conditions to results.

  In various embodiments, the unit operates by using energy from the impact to perform a unit operator after receiving the impact. If a unit has a path to one or more destination units, the remaining energy is directed to one or more destination units. This transmission of energy continues as long as the destination units have a path to their own destination unit and have energy remaining after overcoming energy consumption by the operator and resistance losses on the path. There are many possible energy flow paths in any particular instantiation of the unit model. A particular domain or universe may have different ways in which energy is wasted as the energy flow encounters resistance.

  Referring to FIG. 6, a flow chart illustrates the handling of impacts within a single unit, according to one embodiment of the present invention. FIG. 6 illustrates the processing of a flow in a single unit, according to one embodiment of the present invention. However, certain domains or universes may handle flows differently. Since a practical model comprises a plurality of connected units, it is necessary to repeat the same process through all units. Processing begins at 601 and proceeds to decision 602. Decision 602 determines whether the unit under consideration has received an impact or some energy from another unit. If 602 is false (no), unit processing ends at 611. If 602 is true (yes), processing continues to decision 603.

In some embodiments, the amount of impact energy must exceed a threshold for activating the unit. Decision 603 determines if the impact energy is sufficient to activate the unit. If decision 603 is false (no), unit processing ends at 611. If 603 is true (yes), processing continues to 604.

  At 604, the unit operator is activated. Then, at 605, a loss of consumed energy is applied to determine the remaining energy. In some embodiments, unit activation uses or wastes energy from the input. In other embodiments, no energy is lost due to operator activation and all input energy remains for propagation to the attached unit. In yet other embodiments, energy is added when the unit is activated.

  At 606, resistance adjustment is applied. In some embodiments, activation of a unit reduces its resistance or reduces its threshold, thus increasing the likelihood that the unit will be included in subsequent flows. In other embodiments, activation of a unit increases its threshold or resistance, making it less likely that the unit will be included in future flows.

  When processing is complete at 604 and 605, the unit may have energy to pass to other connected units. At 606, it is determined whether another unit is connected. If the unit is not connected, the process ends at 611. If at least one other unit is connected, decision 606 is true (yes) and processing continues to 607.

  At 607, the remaining energy of the incoming flow is distributed among the output connections. In one embodiment, the energy is equally divided between the output connections. In another embodiment, if there are multiple output connections, the distribution algorithm determines how the energy is split. Distribution algorithms may also constrain flows to paths that have relationships that meet certain criteria. The process ends at 611.

  The unit model has a number of devices that can be activated to cause changes to the information units used in the model. In various embodiments, the method uses such devices to pursue desirable results or avoid undesirable results. A particular domain or universe may have criteria for the desired outcome of the domain or universe itself based on various conditions.

  The previous sections [098] to [0103] described the activity after the unit received an impact. In principle, an unlimited number of units can receive an impact at the same time. In theory, this is completely acceptable. In practice, however, the available hardware dictates how the system manages these impacts. There are numerous means by which this can be achieved. In the case of an in-line system, one embodiment of this system employs a set of impact stacks that are systematically traced from top to bottom by the engine. A unit with an operator that impacts another unit may add impact internally to one of the stacks in this set of stacks. A stack is a structure that is a list of impacts. The impact is a structure having characteristics related to impact energy and path characteristics. A path defines an impact source, a destination, and a relationship that connects the two and flows energy. Stacks, impacts and paths can be implemented by engine or unit models. Impact is defined by the path through which energy flows.

The impact may be added at the end of the current stack, or a new stack may be created where the impact is added, and the new stack is set to the current stack. Thus, the flow to the target unit is completed before the preceding stack resumes its processing. It should be noted that multiple flows can operate in parallel, but serially in time. The way energy is distributed to connections with units is a process that can be implemented by a stack.

  > And! Consider the set of units connected by>. If the truth value of the unit is true, the flow goes> pass, if false,! > Follow the path. > And! > The resistance for all paths other than the relationship is set to infinity. As such, the flow can be guaranteed to follow one of those paths. By installing many units together with similar paths and resistance settings, the flow is directed through a predetermined set of units. These units have operators that perform functions such as mapping, condition determination, unit creation / deletion, unit change, actual interface operation, and dynamic path resistance setting. This set of units is called a process.

  The General Problem Solver of FIG. 8 is an exemplary process that embodies an algorithm used to determine an unknown quantity whose value is determined by an equation. The GPS 801 applies the basic principles of the domain at 802 and 803 to employ all of the necessary calculations to calculate an unknown quantity from a known quantity. The GPS 801 may be designed without using a specific model as a reference, so it can be applied to any domain described by the unit model. The GPS 801 loops on all the principles as described above and applies the principles by impacting the process. If there are no more principles to apply, the loop action follows the path to the “Solve Variable” unit that has a false truth value and begins in box 805. Box 805 loops over a list of equations where each equation is a unit and its elements represent numbers, variables and other arithmetic symbols in a particular sequence. An equation can be created by impacting the unit whose operator is the creation operator. The equation element may be set at that time or later. An equation list is a unit whose elements are lists of equations. The equation is checked in step 807 to see if it contains a solution variable by performing a crossover operation to see if the Solve Variable is a member of the equation (Eq) element. If Solve Variable is a member of the equation element, flow proceeds to the next unit that impacts the process of solving the equation for the variable at step 808. The equation is then evaluated at step 809 and is successfully output at step 8010. If there is no Solve Variable, the process fails at step 806.

  A goal-oriented flow is a flow that achieves a desired result or satisfies a goal criterion. In one embodiment, the goal criteria is represented as a set of start units and a set of end units. In other embodiments, the target start and end units may be represented as situations or states. Various methods can be used to find the path and unit that connects the ending unit to the starting unit within a particular instantiation of the model.

  A critical feature of the unit model is the identification of the units and connections that make up the goal-oriented flow and the flow that satisfies the goal in model instantiation. A related concept is “directed flow”. The directed flow comprises a goal, the domain from which the flow begins, and the resulting flow. Exemplary applications of goal-oriented flows include discovery, general problem solving, and hypothesis testing. A “hypothesis” is the process of developing a new class of knowledge units from known knowledge units.

  In various embodiments, the instantiation of the unit model presents feedback, regulation and observation. As an example, consider two situations denoted S1 and S2. Certain information units are members of both S1 and S2. S2 includes some of the units in S1. Some S1 units are targets of operators of one or more units in S2. Therefore, S2 can monitor S1 and change the content of S1 according to a predetermined standard. This function provides the basis for self-observation and self-regulation through feedback. This concept can be extended to more complex regulatory scenarios involving more than two situations, such as S1-> S2-> S3-> S1.

  The preceding discussion explains the use of models to predict, explain, and pursue what is desirable, based on the general principles of deduction, assuming that correlations are generally causal . However, the unit model has further beneficial advantages. The model can be used to discover knowledge, reflect experience, and generalize.

  The model may also maximize efficiency in finding new knowledge from the experience information or reaching the goal. Maximizing efficiency can be stated in terms of dynamics. As explained, each flow is energy dependent. The possible flow paths may exceed the energy available to generate the flow. Therefore, it is desirable to reduce the energy required to achieve the goal.

  There are many approaches to creating or extending a unit model instance. As discussed, observations can be used to add unit model instances. In some embodiments, it may be desirable to create or extend an instance of the unit model from another type of information rather than from observation. In one embodiment, such conversion is accomplished by application of a general modeling process (GMP) in the dynamics of the system. GMP is a set of methods for dynamically converting knowledge from a common format to a unit model format. GMP is a generalized technique for deploying specific domains.

  Referring to FIG. 7, FIG. 7 shows a flow chart describing one embodiment of a GMP conversion process. The process shown in FIG. 7 is general and can be adapted to the needs, domains, or required data transformations of a particular embodiment. The steps may be performed in a different order, or some steps may be omitted as needed.

  In FIG. 7, at step 701, a domain is identified and relevant data is collected from the real or physical world to determine the scope of model instances to be created. In step 702, a goal for the domain identified in step 701 is identified. Goals can be identified in terms of desirable information units or in terms of one or more processes. In step 703, the required objects are identified, including the class and attribute type identification appropriate for the domain identified in step 701. At 704, unique items in the domain that were not identified in step 703 are identified and captured.

  At step 705, lists and encapsulations are identified, including constraints. At step 706, similar objects are grouped to identify the particular class associated with the class type identified at step 703. Similarly, in step 707, the specific attributes required for each class definition are identified.

In step 708, the principles and conditional relationships are identified and identified according to the defined domain class. In step 709, the required process is identified. Typically, such processes are derived from goals, objects, and classes defined for the domain, but may be associated with class attributes.

  There are many ways to actually implement a unit model. In one embodiment, the unit model is implemented manually using, for example, paper and a drawing device. In any instance of a model that exceeds a small number of units or relationships, it is practical to use a digital computer to store and process the units of the model according to the features and descriptions herein. Thus, in one embodiment, the unit model is provided as a set of data structures, a set of algorithms, and a set of programmed functions that allow the digital computer to store and execute the model. . In another embodiment, an analog computer is used. In another embodiment, a hybrid computer is used that incorporates both analog and digital components. In one embodiment, hardware components specifically designed for the unit model are built to store or execute the model.

  In one embodiment, the computer implementation includes a visualization application. As described herein, unit models can be used to visually display and organize data. Visualization applications provide the ability to view data on a computer display or printer. In one embodiment, the computer implementation includes an editor. The unit model editor allows the user to retrieve, inspect, modify, add, store and document model elements. The editor is similar to the visualization application, but adds functionality to make changes to the model. Thus, the editor can be used to create, modify, maintain, or delete part or all of the knowledge domain. The editor can be designed to edit the elements of the model using a text description or a graphical display.

  In another embodiment, the computer implementation includes an observation module. The observation module automatically processes the events to form units in the unit model and stores the resulting model. Events can be obtained from any source, for example, from a data file, data network connection, simulator, real time monitor, or from external instrumentation and measurement equipment.

  In one embodiment, the computer implementation includes an operator module for acting on information units in the model. The operator module may perform operations including conditional association, causal identification, value assignment, editing, self-observation, impact initiation, impact resistance, threshold determination, and generalization. An operator module may include a mechanism for discovering and generalizing relationships between elements of data.

In one embodiment, the computer implementation includes a dynamic engine module for coordinating the interaction of other modules and controlling the flow of model elements. In one embodiment, the computer implementation includes circuitry for storing and retrieving models in persistent storage, eg, on magnetic tape, rotating disk, or optical storage. In one embodiment, the implementation includes an interface for storing a model in a database and retrieving the model from the database. Many database programs are available that provide benefits such as reliability, redundancy, and the ability to accommodate large amounts of data. In one embodiment, multiple applications are integrated into an information modeler toolset or workbench. The integrated toolset includes a user interface that allows the user to control, manipulate, and selectively execute the functions of the model.

  There are many computer applications and tools that can be implemented in accordance with the features and descriptions herein. The exemplary embodiments herein are merely examples and should not be considered as comprehensive or limiting. In one embodiment of the features and descriptions herein, the computer implementation of the unit model is a set of computer functions, a set of objects, a set of subroutines, or a module whose features can be invoked by an external computer program. Offered as a set. This allows the user to embed the unit model implementation in another computer program without coding all the functions required to implement unit model access, operation and execution. To do.

  The features and descriptions herein are useful in modeling a wide variety of domains. Several exemplary embodiments and instantiations of application examples of features and descriptions herein are illustrated in the figures and associated descriptions. It should be understood that not all unit model features need to be employed in modeling a particular domain. Instead, the model provides a rich set of features and options that can be employed as needed. Similarly, some aspects of model functionality can be adapted to match a particular domain without changes to the model itself.

  For example, in the resistance, dynamics, and flow aspects, each domain model may require different features. In some embodiments, the energy of the flow is stored such that the sum of the energy output and wasted energy is equal to the energy input to the system. In other embodiments, some units regenerate or amplify energy during propagation. The particular domain or universe determines the rules that apply to resistance, dynamics and flow.

  Referring now to FIG. 3A, there is illustrated one embodiment of the present invention applied to the kinematic domain of point objects in a closed system. This illustrates the unit model function for capturing physical system features. At 302, the highest system class is shown. Class 302 has various attributes. Not all attributes are necessarily shown in the figure. Exemplary attributes of class 302 are shown as name attribute 303, object_list attribute 304, and coordination system attribute 305. Class 302 is also associated with process list 306.

  In some embodiments, an attribute that is an encapsulation is associated with a type unit, a class unit, and a constraint list (“CList”) unit. Thus, with further reference to FIG. 3A, the object_list attribute 304 is associated with the object_list type atom 307, the object_list class atom 308, and the object_list CList atom 309. Class atom 308 may be associated with one or more class instances. Point object class 310 is one such instance associated with class 308.

Referring now to FIG. 3B, the example from FIG. 3A continues. The point object class 310 from FIG. 3A is shown with exemplary associated attributes. Attributes include name 320, displacement function 321, velocity function 322, acceleration function 323, instance 324, and transition 328. The point object class 310 is also associated with the process list 332. Each of these attributes and processes adds properties or functionality to the point object model defined by the point object class 310.

  Instance attribute 324 is an encapsulation of the instance of the point object in the model in this exemplary embodiment. Instance attribute 324 is associated with instance type 325, instance class 326, and instance CList 327. Similarly, transition attribute 328 is associated with transition type 329, transition class 330, and transition CList 331.

  Point object class 310 is associated with various processes through process list 332. Two exemplary processes are shown. Process 333 defines a process for calculating velocity as a time derivative of displacement, and process 334 defines a process for calculating acceleration as a time derivative of velocity.

  An exemplary embodiment of point object kinematics continues in FIG. 3C. Transition class 330 (from FIG. 3B) is shown with an instantiated point object transition class 340. Some associated attributes of class 340 are shown. The attributes include an initial instance 341, a final instance 342, a displacement change 343, a time change 344, a speed change 349, and an acceleration change 350. Class 340 is associated with process list 345. Process list 345 is associated with three processes. Process 346 defines a process for calculating displacement changes. Process 347 defines the calculation of speed change. Process 348 defines the calculation of acceleration change.

  An example of point object kinematics continues in FIG. 3D. In FIG. 3D, the instance class 326 from FIG. 3B is further described from a model perspective. Instances of class 326 are shown as point object instances. Several attributes of instance 360 are shown including label 361, time 362, velocity 363, acceleration 364, and displacement 365. Instance 360 also includes a process list 369. The displacement attribute 365 is an encapsulation having an associated atom displacement type 366, an atom displacement class 367 and an atom displacement constraint list 368.

  An exemplary embodiment of point object kinematics continues in FIG. 3E. The displacement class atom 367 from FIG. 3D is further described from a model perspective. Vector class 370 is an element of displacement class 367. Vector class 370 is associated with attribute adjustment system 371, attribute Cartesian 372, attribute pole 373, and attribute sphere 373. Vector class 370 is also associated with its process list 375.

  Further details of an exemplary point object kinematics embodiment are shown in FIG. 3F. 3F shows the Cartesian attribute 372 components and associations, and FIG. 3G shows the polar attribute 373 components and associations, both from FIG. 3E. In FIG. 3F, a Cartesian attribute 372 is associated with a Cartesian class 380 that is associated with a Cartesian vector class 381. The Cartesian vector class 381 is associated with the process list 385 and the X attribute 382, Y attribute 383 and Z attribute 384. In FIG. 3G, pole attribute 373 is associated with pole class atom 390 associated with pole vector class 391. The pole vector class 391 is associated with an R attribute 392, a theta attribute 393, and a Z attribute 394.

Here, one approach for making a discovery is described. FIG. 4A-1 illustrates the class structure for a company. This structure tracks data about various aspects of the company, such as purchases, contracts and sales. We want to know if there is a correlation between one set of data and another set of data. Such correlation is commonly referred to as association rules and is a standard structure in knowledge discovery and data mining.

  4A-1 and 4A-2, one embodiment of the present invention is illustrated as applied to a domain of business operations. This illustrates the ability of the unit model to represent entities such as people or companies. This also illustrates the representation of abstract relationships between entities such as employers / employees and contracts. Two top classes are shown, class company 400 in FIG. 4A-1 and class company sales 450 in FIG. 4A-2. Class company 400 has attributes that are useful in modeling exemplary domains. The attribute name 401, attribute purchase 402, attribute contract 403, attribute employee 404, and attribute rental 405 are all aspects of the class company 400. Company class 400 is also associated with process list 406, which includes exemplary process expense schedule creation 407. Procell 407 is further detailed in FIG. 4F.

  With further reference to FIG. 4A-2, company sales class 450 is associated with attribute name 451 and attribute predicted sales 452. Class 450 is also associated with process list 453. 4B, 4C, 4D, and 4E provide further details of example business operations through extending the definition of selected elements introduced in FIGS. 4A-1 and 4A-2.

  FIG. 4B provides details of the purchase attributes 402 defined in FIG. 4A-1. Purchase attribute 402 is associated with purchase type 410, purchase class 411, and purchase CList 412. Purchase class 411 is associated with purchase class 413 having attributes of description 414 and payment 415.

  FIG. 4C provides details of the contract attributes 403 defined in FIG. 4A-1. The contract attribute 403 is associated with the contract class atom 420. Contract class atom 420 is associated with service contract class 421, which has attributes of description 422, start date 423, monthly amount 424, and payment 425. The class service contract 421 is also associated with the process list 426.

  FIG. 4D provides details of the employee attributes 404 defined in FIG. 4A-1. Employee attribute 404 is also associated with employee class atom 430. Employee class atom 430 is associated with employee class 431, which has attributes of name 432, start date 433, salary 434, and payment 435. Class employee 431 is also associated with process list 436.

  FIG. 4E provides details of the forecast sales attribute 452 defined in FIG. 4A-2. The predicted sales attribute 452 is associated with the predicted sales class atom 460. The predicted sales class atom 460 is associated with the predicted sales class 461. The predicted sales class 461 includes an explanation 462, a unit number 463, a start date 464, an end date 465, a unit price 466, a net amount 467, and a payment 468. Has attributes. The class forecast sales 461 is also associated with the process list 469.

Referring now to FIG. 4F, details of the expense schedule creation process 407 from FIG. 4A-1 are provided. Process 407 consists of a sequence of steps or operations, and each step of the process may itself be a sequence of operations. Thus, the unit model can be used to describe complex processes, layered processes, and multilayer processes. Referring to FIG. 4F, the process 407 comprises five steps 470, 471, 472, 473, and 474. Next, the process 470 comprises three steps 475, 476, and 477. Process 471 comprises two steps 478 and 479. The process may also include detailed actions, loops, or conditional decisions. Process 475 includes loop 480. Process 476 includes loop 481. Process 477 includes loop 482.

  FIG. 11 defines process 477 of FIG. 4F. The process has four parameters at step 1101. The variable [CurrentDate] is defined in step 1102 and set in the [BeginDate] parameter in step 1103. If [CurrentDate] is less than [EndDate] at step 1104, then at step 1105, an instance of the payment class is created. The new class instance is the target of the copy unit. In step 1106, the target is identified as {New Payment}. Braces indicate that the element refers to another unit's target. The attributes of the payment instance are set at step 1107, [CurrentDate] is incremented by 30 days at step 1108, and the payment instance is added to [Payment] at step 1109.

  5A-5E illustrate one embodiment of the invention applied to the domain of object identification in an array of black and white pixels. This illustrates the model's ability to find patterns and relationships in the data. In FIG. 5A, a Cartesian pixel array class 501 exists at the top and comprises a two-dimensional array of picture elements or pixels. Array 501 has several associated attributes including pixel attribute 502, row attribute 503, row_count attribute 504, column attribute 505, column_count attribute 506, blob attribute 507, shape attribute 508, and object attribute 509. A process list 510 is also associated with class 501.

  FIG. 5B-1 provides further details of the pixel attributes 502 already defined in FIG. 5A. Pixel attribute 502 is associated with pixel class atom 520. Pixel class atom 520 is associated with pixel class 521, and pixel class 521 has attributes of row 522, column 523, and value 524.

  FIG. 5B-2 provides further details of the row attributes 503 already defined in FIG. 5A. A row attribute 503 is associated with the row class atom 525. Row class atom 525 is associated with row class 526, which has attributes of pixel 527 and position 528. Any class in model instantiation may have many attributes and subclasses not shown. Similarly, attributes may be added or changed as part of adjusting, expanding, or maintaining model instances, or in response to changes in the domain being modeled. For example, details of row attribute 503 are shown, but column attribute 505 has similar attributes. Accordingly, the details of the exemplary embodiments are intended to be exemplary of model functionality rather than being comprehensive.

FIG. 5C provides further details of the blob attribute 507 already defined in FIG. 5A. Blob attribute 507 is associated with blob class atom 530. Blob class atom 530 is associated with blob class 531, which has attributes of absolute pixel 532, difference pixel 533, sequence 534, line 535, type 536, and shape 537. Blob class 531 also has an associated process list 538.

  FIG. 5D illustrates the process used by class 501. In FIG. 5A, class 501 has been shown to be associated with process list 510. In FIG. 5D, several processes that are elements of process list 510 are shown. Process 540 creates a pixel array. Process 541 shows the pixel array as a button array. Process 542 obtains the selected button. Process 543 creates a blob. Process 544 maps the basic shape. Process 545 maps complex shapes.

  FIG. 5E illustrates the process used by blob class 531. In FIG. 5C, class 531 has been shown as being associated with process list 538. In FIG. 5E, several processes that are elements of process list 531 are shown. Processes 550, 551, 552, 553, and 554 perform the processes required by class 531.

  The model can generally be designed to account for a large number of actual / observed situations. The combined sum of all actual situations that are actually explained in the model is called the data domain. A relationship can exist between a certain set of data and other sets of data, perhaps all such sets of data have values in some numerical range. A unit or set of units can be identified as a source and target domain. Finding a set of all theoretically possible map steps that link a source to a target unit is a simple process for those skilled in the art. As such, embodiments of the system of the present invention can autonomously discover new knowledge. Therefore, it may be required for the unit model to discover these principles, and even the classes and attributes to which the principles apply. This process is called discovery. Other common names exist for this process, including knowledge discovery and data mining.

  The flow chart diagram of FIG. 9 illustrates a possible implementation of the discovery process. The discovery process begins by impacting the process discovery (step 901). In step 902, the process of creating a list of units representing all of the paths describing the domain is shocked. The element of each unit is a [source] relationship [destination]. The process is then impacted at step 903 to create a set of maps (maps: _1, _2, _3). The process may create a map for each potential possibility, for example, _1 relationship 1_2, [source] _1_2. Map creation 903 defines a process for creating a map of the form _ 1 relationship _ 2 for every concept in the concept list. The “_1_2_2_3” map is applied to each concept, then a new unit is created and its elements are! _1 _2! Set to _3. this! Indicates that the element is defined as _1 for using the content of _1. The mapping process is implemented as a new map that is added to the existing map in step 1005, with each concept in step 1003 resulting in a new reference to each concept in step 1004. FIG. 10 shows a process 1001 for creating a map by looping over a list of. Returning to FIG. 9, at step 904, the process of bringing together all of the map combinations or a subset of all of the map combinations into associations of precedents and results is impacted. Rule classes with precedent, result and metric attributes are copied. In step 905, the process is impacted to test the rules. The test process can track any set of metrics for a rule and store the results in the attributes of the metric attribute of the rule class instance.

  It is easy to figure out how the described process variations make them more general and complex. For example, the number of rule precedents can be much greater than 1, and the combinations of virtual units in the map can be more general.

  In addition to a particular domain or universe, it is also anticipated that the system of the present invention can be implemented without restrictions for a particular domain. Thus, in one embodiment, an object is instantiated that includes an encapsulation of all situations experienced and an encapsulation of all pleasure and pain values. This object is referred to as “self”.

  In one embodiment, a situation called S_self is observed that observes the Self object and regulates the Self object towards the goal. In one embodiment, the goal is to seek joy and avoid pain. This goal can be observed to reflect the primary drivers of conscious organisms. With further improvements including the addition of stimuli and possibly other goals, more complex behaviors can be generated. Therefore, this embodiment allows further study of the nature and operation of artificial intelligence.

  While certain embodiments of the present invention have been described above, it will be understood that the described embodiments are merely examples. Accordingly, the inventions should not be limited based on the described embodiments. Rather, the scope of the invention described herein should be limited only in light of the following claims when considered in conjunction with the above description and the accompanying drawings.

Claims (23)

A system for knowledge representation and application development,
A processor;
A memory unit coupled to the processor;
A plurality of information units stored in the memory unit,
An information unit element or set of information unit elements that directly or indirectly references at least one other information unit;
An operator defining one of a plurality of actions that the information unit performs when the information unit is impacted, ie activated, and
An information unit comprising at least one path describing a relationship existing between the information unit and another information unit;
Impact the operator of an information unit, ie activate the operator, and further impact, ie activate, another information unit to which the impacted information unit is connected via a path A dynamic engine module executable by the processor,
The information unit and the path embody a knowledge representation schema, an information unit model, whereby an instantiation or instance of the information unit model is a domain, a set of observations, actual knowledge, and data A system that references a specific set and assembly of model elements to correspond to at least one of the following. A system for knowledge representation and application development,
A processor;
A memory unit coupled to the processor;
A plurality of information units stored in the memory unit,
An operator defining one of a plurality of actions that the information unit performs when the information unit is impacted, ie activated, and
An information unit comprising at least one path describing a relationship existing between the information unit and another information unit;
Impacting or activating the operator of an information unit and further impacting, i.e., activating, another information unit to which the impacted information unit is connected via a path A dynamic engine module executable by the processor to stream,
The information unit and the path embody a knowledge representation schema, an information unit model, whereby an instantiation or instance of the information unit model is a domain, a set of observations, actual knowledge, and data A system that references a specific set and assembly of model elements to correspond to at least one of the following.   3. The system of claim 2, wherein at least one of the information units includes an information unit element or set of information unit elements that directly or indirectly references at least one other information unit.   An object or set of objects that have similar characteristics at some level of abstraction is represented by a set of information units that includes an unlimited number of information units to form a class, and features of a class are represented by multiple attributes 4. The system according to any one of claims 1 to 3, wherein:   The system of claim 4, wherein each attribute of the class represents one of a text string, a list, and a class.   The system of claim 4, wherein the original class is copied to form a class instance having a path to the original class.   The system according to claim 1, wherein the information unit element refers to a set of information units representing a list.   4. A system according to any one of the preceding claims, wherein a set of information units having a path to each other directing one or more shocks or activations to a predetermined set of information units forms a process.   4. A plurality of information units are referred to as an encapsulation having a single information unit, and the resulting encapsulation is one of a list, a process, and a class. The system described in the section.   4. A system according to any one of the preceding claims, wherein one information unit element is a symbol element or symbol sequence for use in a mapping operation.   4. Each information unit includes a characteristic indicating a state of the relationship of the information unit to another information unit when the information unit is impacted, i.e. activated. The system described in.   The impact of the operator, i.e. the activation, allows a selective start of a dynamic characteristic or flow of energy through a predetermined set of information units connected by the path. The system as described in any one of thru | or 3.   Each path has a resistance characteristic that controls the flow of energy from one information unit to another information unit when the other information unit is impacted, ie activated. The described system.   13. Each information unit has a threshold characteristic that impacts the operator, i.e., quantifies the number, amount, or both of the impacts required to activate the operator. System.   The system of claim 12, wherein the flow of energy is directed through a single path to a single destination unit.   The system of claim 12, wherein the flow of energy is directed to a plurality of destination information units through a plurality of paths.   13. The system of claim 12, wherein the energy flow achieves a desired result or meets certain target criteria.   User to enable the user to control, manipulate and selectively execute the functions of the unit model to build a model of a selected area of knowledge by computer implementation of the unit model 4. A system according to any one of the preceding claims, wherein a dynamic engine and a plurality of integrated modules are implemented, including an interface.   The system of claim 18, further comprising a visualization module that graphically displays the unit model.   The system of claim 18, further comprising an interface for storing a unit model in a database and allowing a user to retrieve the model from the database.   The system of claim 18, further comprising an interface to allow a user to incorporate the unit model into an external computer program.   4. At least one of the information units includes a truth value that includes one of a binary or non-binary real value that represents a characteristic of the information unit. The system described in.   The at least one of the information units includes a desirability factor that includes one of an integer or real-valued property that corresponds to desirability of one information unit for other information units. The system as described in any one of thru | or 3.