Using graph patterns to enhance the integration of models into semantic frameworks

Abstract

A computer system (800) is provided, comprising at least one processor (805) for modeling operations (600) related to capturing domain knowledge (100). These operations include, via the processor (805), creating (702) graphical models (400, 500, 600) of inputs to equations (101, 202) related to the domain knowledge (100). The graphical models (400, 500, 600) associate at least one of the inputs with another; and wherein the graphical models (400, 500, 600) associate inputs with outputs (704). The operation (600) also includes deriving (706) enhancement type information from the graphical models (400, 500, 600) and adding (708) the derived enhancement type information to equations (5-8) via the processor (805), the addition contributing to the use of equations (5-8) by artificial intelligence.

Description

I. Technical Field

[0001] This invention relates to capturing domain knowledge for use in artificial intelligence. II. Background Technology

[0002] A vast amount of scientific knowledge can be found in books, journal articles, conference reports, and perhaps most explicitly, in the computer code that encapsulates scientific computing. To some extent, the target audience for all these repositories is people with a certain level of skill in the field. For documents, this requires the skill of correctly interpreting the meaning of the text. For computer programs, this skill is providing correct input—input that is consistent in some way implicitly, or input that is consistent with the document the user must read.

[0003] When such computational models are integrated into the semantic framework of artificial intelligence (AI), a problem arises: AI then expects to leverage these models, linking them together so that the output of one model is the input of another, and so on. In other words, all the implicit knowledge that subject matter experts (SMEs) use to correctly invoke computational models must now be explicitly captured in the knowledge base of AI-based computing devices.

[0004] According to some embodiments, knowledge that is typically implicit (and provided only by, for example, human experts) is explicitly included in the captured knowledge and can be used for machine reasoning.

[0005] For any type of mathematical and / or scientific equation, the implementation allows the use of a graphical model to capture the relationships between inputs and outputs, i.e., domain knowledge of the relationships between all inputs and each other, and between the output and all inputs.

[0006] This refers to any domain where a computational model has been constructed, where appropriate inputs need to be selected and the output needs to be known in domain terminology. Embodiments of the present invention enable AI-based computing devices to apply the knowledge represented in the equations in the same way that can be done by a person skilled in the art.

[0007] Features of some embodiments will be described in the context of illustrative but non-limiting examples. In these examples, a set of equations will be considered to capture the science of flight. These illustrative examples are derived from a website owned by NASA Glenn Research Center. III. Description of the Drawings

[0008] Figure 1 provides a conventional example of a simple domain. This domain relates to the speed of sound. Figure 1 specifically relates to Mach number 100, which is the ratio of the speed of a flying object to the speed of sound. In Figure 1, the aircraft 102 is depicted at different speeds relative to the Mach number. In the conventional illustration of Figure 1, the embodiments of the present invention aim to overcome the following problems.

[0009] Is the speed of sound in Figure 1 the speed of sound in water? Is it the speed of sound in the Arctic air when an object is flying at the equator? While the answer to these questions is no, this answer is not explicitly stated in the equations. That is, the answer to this question depends on one's own knowledge that the object's velocity must be divided by the speed of sound at the object's location in the air. Therefore, if an object is moving in the air at an altitude of 32,000 feet, then the required speed of sound is: at 32,000 feet. SME will understand this assumption when using the equations in Figure 1. AI-based computing devices will not be able to make this assumption and effectively use the simple equations in Figure 1.

[0010] In a more complex example, consider the maximum total thrust of the aircraft as: [Number of engines] x [Total net thrust / number of engines]. How can AI-based computing devices use this equation? The relationship is that the total thrust of the aircraft equals the number of engines multiplied by the net thrust of each engine. It is assumed that all engines on the aircraft have the same net thrust.

[0011] The challenge when attempting to solve the equations is to explicitly capture the relationships between the inputs, and in domain terms, by binding all the inputs and outputs back to the same object. In this more complex example, the object is an aircraft. In the previous example, the object was the atmosphere. That is, an assumption is added to the first example that it is traveling at the speed of sound through a gas. Adding this assumption means that the equations in Figure 1 are valid only if the medium is a gas (e.g., the atmosphere).

[0012] Therefore, the speed of sound in the denominator of Figure 1 is the speed of sound in the air through which the object passes. Similarly, the fact that the velocities in the numerator and denominator must be in the same units is not explicitly stated in the equation or the text.

[0013] The equation for the speed of sound, also provided by NASA's Glenn Research Center, is shown in the conventional illustration 200 of Figure 2.

[0014] More specifically, in Figure 200, the specific heat ratio γ in Equation 202 can also be calculated. If the velocity of the object is large, the specific heat ratio is given by the following formula:

[0015] gam=1+(gamma-1) / (1+(gamma-1)*[(theta / T)^2*e^(theta / T) / (e^(theta / T)-1)^2])

[0016] in

[0017] gamma (γ) is the specific heat ratio of a calorimetric perfect gas, which is 1.4 for air. theta is the thermal constant, which is 5500 degrees Rankine or 3056 degrees Kelvin. T is the static temperature.

[0018] However, static temperature can be calculated from the altitude of a "standard Earth day," as shown below. If the altitude (h) is less than 36,152 feet (troposphere), the temperature (Fahrenheit) is given by the following formula:

[0019] T = 59 - 00356h

[0020] If the altitude h is between 36,152 and 82,345 feet, then the temperature is:

[0021] T = -70

[0022] If the altitude h is greater than 82,345 feet, then the temperature is:

[0023] T = -205.05 + 00164h

[0024] In order to use these equations correctly, AI-based computing devices must know how the inputs and outputs of each equation are related in domain terminology, and must have explicit knowledge of the input and output units of each equation, and ensure that they are aligned when using multiple equations.

[0025] Figure 3 is an illustration of the conventional user interface 300. This implementation simplifies the task for human users at NASA's Glenn Research Center by wrapping the computational model in a Java applet. The applet allows users to input numerical values ​​302, which can then be used to calculate the values.

[0026] This method requires the user to select a unit. As shown in Figure 2, the desired unit implied by this selection is displayed in the applet interface. The Java applet code implements the task of linking various computational models together in a consistent manner, incorporating appropriate assumptions into the implementation.

[0027] Figure 3 shows the interface of a mini-program from NASA's website, where users can input numbers and use equations to calculate values, such as speed. Users can input altitude and the speed of the spacecraft; "speed is the input."

[0028] In the mini-program interface of Figure 3, all domain knowledge is captured in the user interface. However, the mini-program is still not constructed in a way that enables AI-based computing devices to interpret and fully analyze the units of input and output to intelligently apply the equations in Figure 2. At best, these types of mini-programs merely make the input of data values ​​a more consistent process for human users.

[0029] In an exemplary embodiment of the present invention, one method for capturing and transmitting knowledge of a domain includes understanding the relationships between the values ​​of various inputs to equations representing the domain, and understanding the relationships between those values ​​and the outputs of those equations. Computational modeling is a method that can define these relationships.

[0030] There are various formal modeling methods that can be used to capture and transfer knowledge of a domain. Predicate logic is a very powerful form of formalism and is frequently used. If predicate logic is limited to predicates with two or fewer elements, it corresponds perfectly to the graphical approach used by the Web Ontology Language (OWL). This is because predicates with two elements are edges in a directed graph. When using graphical representations, predicates with more than two elements must be represented by creating additional mediating concepts that map the arguments of predicates with higher elements together. Concept graphs (CG), ISO standard common logic (CL), and knowledge exchange formats (KIF) are suitable alternatives within the spirit and scope of this embodiment.

[0031] For our example above, we can use Figure 4-6 The graphical model shown is used to represent the Mach number field. Specifically, Figure 4-6 The graph modeling in this paper explicitly captures several important relationships within the Mach number domain and provides illustrative examples of how to use ontology to describe graph relationships.

[0032] exist Figure 4 In the example, Figure 4 This is an exemplary diagram of a simple class hierarchy 400 for Gas 406. The class hierarchy 400 serves as a simple classifier to accumulate the attributes of PhysicalThing 402. Figure 4 In the class hierarchy 400, physical objects 402 are included, with matter 404 being modeled as a subclass. Gas 406 is modeled as a subclass of matter 404, and air 408 is modeled as a subclass of gas 406.

[0033] Figure 5 This is an exemplary diagram of another simple class hierarchy 500 for Physical Objects 502. Class hierarchy 500 shows the existence of classes such as... Figure 4Physical object 502 is a subclass of physical object 402. Simply put, it is a specific physical object 502 that can have a Mach number (i.e., is a subclass of physical object 402) and move through the air at a Mach number 408. Figure 5 In this context, physical object 502 is constrained to a single, normalized, unitless Mach number 504. (Note that if a temporal model is being captured, the Mach number will be a single value at any given moment.)

[0034] Figure 6 Figure 600 is an exemplary diagram of a related subclass that is part of the Mach domain, and the scope of physical object 502. In particular, Figure 600 explicitly captures several important relationships related to physical object 502 and gas 406.

[0035] For example, gas 406 has a molar mass of 614. This value works because the gas constant 616 can be calculated from the general gas constant and molecular weight, but the molar mass 614 of the gas being calculated must be known. As an example and not a limitation, physical object 502 and gas 406 include subclasses:

[0036] The physical object (502) has many important properties in the domain, including:

[0037] Speed ​​(velocity) 602

[0038] Force 604 (Important for thrust, flight model)

[0039] movesIn 606, with range of gases

[0040] Mach number 608, range in decimal.

[0041] Mass 610 (Important for thrust, flight model)

[0042] Temperature 612 is inherited from physical phenomena; gases also inherit this value.

[0043] Gas (406) includes several important properties associated with gases. For example, these properties include:

[0044] The molar mass (molarMass) is 614, used to calculate the gas constant (not shown in the above equations for simplicity).

[0045] The gas constant is 616, which is shown in the equation above.

[0046] gamma 618, calorimetric specific heat ratio of a complete gas

[0047] gammaCal 620, specific heat ratio of incompletely calorimetric gases

[0048] Speed ​​of sound (622), an important property in the Mach equations mentioned above.

[0049] Drawing a domain model explicitly captures the relationships between various classes whose attribute values ​​are inputs and outputs of the aforementioned equations. The current challenge is how best to correlate this domain information with the computational model represented by these equations.

[0050] Most programming languages ​​have the concept of built-in types, such as integers, floating-point numbers, strings, booleans, etc. These types are used to specify the type of a variable, the signature of a method call, and the type of a value or return value. In object-oriented languages, classes can be defined that align with classes in a domain model, and these classes can then be used as types. However, there are significant differences between the expressiveness of most object-oriented languages ​​and that of graph-based ontology languages. Most importantly, in most object-oriented languages, attributes are represented as fields within a class and do not exist independently. In contrast, attributes in graph ontology languages ​​are first-class citizens and can be restricted based on the class of the thing with the attribute, such as value type, cardinality, etc. This means that multiple classes can reside in the domain of the same attribute, and attributes can be restricted in cardinality, value type, etc., which differ between different classes but are still identifiable as the same attribute. Therefore, even in the case of object-oriented languages, the useful part of the code may be methods of built-in or primitive types. This is, for example, the Java applet on NASA Glenn Research Center's hypersonic website. Aside from the applet itself, the only non-built-in class used is the user interface class.

[0051] When integrating computational models whose inputs and outputs are typed according to the language's built-in data types into a semantic framework, both built-in data types and semantic domain types are important. However, these two types are not sufficient.

[0052] Consider some examples of using the equations above. These examples will be illustrated in the Semantic Application Design Language (SADL), but the language may differ. The key points are the information content—(1) the primitive data types of the inputs and outputs, (2) the domain concept that the inputs or outputs are values, and (3) how the inputs and outputs relate in domain terms. Below are the characteristics of the four equations above, the speed of sound, the gamma of an incompletely calorimetric gas, the atmospheric temperature according to altitude, and the Mach number of a flying object, all of which are inputs and return values ​​based on the language's built-in types.

[0053] In Equation 1-8, the input is: TGRQ. In this case, just for fun, we input the unit "us", which is the unit system specification, and it returns double.

[0054] (T) is the temperature of the gas, expressed in Kelvin or Rankine.

[0055] (G) is the gamma of "the" gas. They use the indefinite article, just as in English.

[0056] (R) is the gas constant of "the" gas. If it is a metric system (consistent with Kelvin), it must be "g / mole"; if it is an imperial system, it must be "lbm / lbmole".

[0057] (Q) is theta, a constant expressed in Kelvin or Rankine. There are two examples of the unit theta: one for the metric system and one for the imperial system. The correct system must be used with the correct unit.

[0058] (us) indicates the unit system, either metric or imperial. If it is metric, then the first group of units is best used in all enhanced types. If the imperial system is applicable, then use the second group of units.

[0059] This equation determines the speed of sound for the same gas, in meters per second or feet per second.

[0060] More specifically, Equations 1-4 below do not contain any enhanced type information.

[0061] External CAL_SOS(double T,double G,double R,double Q)returns double.

[0062] External CAL_GAM(double T,double G,double Q)returns.double

[0063] External tempFromAltitude(double alt)returns double.

[0064] External computeMach(double alt,double R,double G,double Q,

[0065] `double vel) returns double`.

[0066] It is also essential to explicitly capture how the inputs and outputs of the computational model are interconnected within the domain model. As illustrated in the diagram above, these relationships can be associated with the computational model as a means of making these relationships explicit. For any single computational model, the scope of the desired graphical model is a domain subgraph that connects all inputs and outputs together.

[0067] Using the exemplary equations from Figure 2, add the domain model graphical patterns necessary to capture the explicit context of the equations, such as... Figure 6 The graphical model 600 is shown. Furthermore, units of measurement supported by the graphical model 600 are provided. More specifically, equations 5-8 include enhancements that enable AI-based computing devices to explicitly capture the appropriate use of the equations in Figure 2.

[0068] This invention relates to the input of an equation: the velocity of an object / speed of sound. These are correlated through a domain model. In other words, it is the speed of sound of the object (the numerator), and the denominator is the speed of sound "as the object moves through the air".

[0069] This invention allows for the explicit capture of that type of domain knowledge so that artificial intelligence can determine how to use the equation “Mach number equals the speed of an object / speed of sound”.

[0070] Figure 6 The combination of (ontology graphs) defines the ontology graphically. This is the ontology used to eliminate ambiguity in the meaning of these equations (5-8).

[0071]

[0072]

[0073] In Equation 5, the graphical model only needs to relate its values ​​to the properties of a specific instance of a gas class as input and output. In this syntax, "a Gas" is used in the first parameter, "the Gas" in subsequent parameters, and the return type explicitly indicates that all refer to the same Gas. Therefore, this use of structured English language is consistent with its meaning in standard English usage. The same applies to Equation 6. In Equation 7, the computational model described is actually for air, not just any subclass of gas. This is important for capture.

[0074] In Equation 8, which illustrates a more complex equation involving multiple concepts, the inputs and outputs are correlated via a graphical pattern that rises to nodes representing instances of PhysicalObjects (defined as "a PhysicalObject" in the first parameter and "the PhysicalObject" in subsequent parameters and return values). Note that the first reference to "Air" is "some Air." This is a permissible alternative to "an Air" because it sounds more natural for matter. Also note that the equation may reference multiple PhysicalObjects, in which case the identity of each distinct object must be explicitly clear.

[0075] This can be achieved in the SADL language by using "a PhysicalObject", "the PhysicalObject", etc. for the first, and "a second PhysicalObject", "the second PhysicalObject", etc. for the second. This extends to "third PhysicalObject", etc. (Inventory of Concept-level Rules, U.S. Patent Reference No. 317709-US-1, and the paper Concept-level Rules for Capturing Domain Knowledge, A. Moitra, A. Crapo, R. Palla. 12th IEEE International Conference on Semantic Computing, January 31-February 2, 2018. https: / / ieeexplore.ieee.org / abstract / document / 8334469 / , covering this usage of indefinite and definite articles, the contents of which are incorporated herein by reference in their entirety for all purposes).

[0076] Other important information to capture is the constraint on input and output units. Equation 7 shows the case where only a single unit set applies: "ft" as the input unit and "[degree] Rankine" as the output unit. Other equations are slightly modified from those shown in Equations 1, 2, and 4 to obtain additional parameters indicating whether the unit system is metric or imperial. For example, in Equation 5, if the UnitSystem is metric, the first parameter must be in Kelvin, the third parameter in "g / mole", the fourth parameter in Kelvin, and the return value in "m / sec". Similarly, if the UnitSystem is imperial, the units are Rankine, "lbm / lbmole", Rankine, and "ft / sec". Note that the second parameter is unitless. The presence of the unit system parameter implies that the equations are encoded with computations using two unit systems, with appropriate flow control.

[0077] Figure 8 The device includes a computer system 800. The computer system may include a processor 805 operatively coupled to a communication device 820, a data storage device 830, one or more input devices 810, one or more output devices 820, and a memory 825. The communication device 815 may facilitate communication with external devices such as reporting clients or data storage devices. The input device 810 may include, for example, a keyboard, buttons, a mouse or other indicating device, a microphone, a knob or switch, an infrared (IR) port, a docking station, and / or a touchscreen. For example, information may be input into the device 800 using the input device 810. The output device 820 may include, for example, a display (e.g., a screen), speakers, and / or a printer.

[0078] Data storage device 830 may include any suitable persistent storage device, including combinations of magnetic storage devices (e.g., magnetic tape, hard disk drives, and flash memory), optical storage devices, read-only memory (ROM) devices, etc., while memory 825 may include random access memory (RAM), storage class memory (SCM), or any other fast access memory.

[0079] Service 835 and application 840 may include program code executed by processor 805 to cause device 800 to perform any or more of the processes described herein (e.g., FIG. 2). Embodiments are not limited to the execution of these processes by a single device.

[0080] Data 845 (cached or complete database) may be stored in volatile memory (e.g., memory 825). Data storage device 830 may also store data and other program code and instructions, such as device drivers, operating system files, etc., necessary to provide additional functions of device 800 and / or to operate device 800.

[0081] Service 835 and application 840, including one or more processing modules (e.g., enhancement type module 840a), perform specialized processing to enhance the integration of the model into the semantic framework. Processing modules #2 (840a) and #3 (840b) may include program code executed by processor 805 to cause device 800 to perform any one or more of the processing described herein (e.g., Figure 4-7 The embodiments are not limited to these processes being performed by a single device.

[0082] IV. Detailed Description of the Background

[0083] The problem this invention addresses is that much scientific knowledge is implicit in people's minds, such as "Mach number #". The very simple equation for Mach number # is: the speed of an object / the speed of sound.

[0084] Is this the speed of sound in water? Is this the speed of sound through the air at the Arctic Circle while the object is flying at the equator? No, but that's not explicitly stated in the equation. It depends on what you know in your mind—that you have to divide by the speed of sound at the object's location in the air.

[0085] So if an object is moving through the air at an altitude of 32,000 feet, then you need the speed of sound at 32,000 feet.

[0086] This invention allows you to explicitly capture that type of domain knowledge so that artificial intelligence can determine how to use the equation “Mach number equals the speed of an object / speed of sound”.

[0087] This invention relates to the input of an equation: the velocity of an object / speed of sound. These are correlated through a domain model. So in other words, it's the speed of sound of the object (the numerator), and the denominator is the speed of sound "as the object moves through the air". This will be a simple example.

[0088] A more complex example: We can say that the maximum total thrust of an aircraft is (number of engines) x (total net thrust / number of engines). How does artificial intelligence use this equation?

[0089] The relationship is that the total thrust of an aircraft equals the number of engines multiplied by the net thrust of each engine. It is assumed that all engines on the aircraft have the same net thrust.

[0090] The problem being addressed is to explicitly capture the relationships between inputs, and between inputs and outputs in domain terminology, by binding all inputs and outputs back to the same object.

[0091] In this example, it is an aircraft.

[0092] In another example, it is essentially the atmosphere.

[0093] Let's make an assumption. In the first case, the speed of sound is through a gas. Therefore, you know this equation only applies when the medium is a gas. So you know it only works for gases, such as the atmosphere.

[0094] Existing technology. The same concept can be applied to data tables, so that if you have tabular data, you can know the relationships between data elements or column headers in any particular row. (Existing technology in this field).

[0095] If we turn to page 2, which contains text from NASA's website, we have an equation that isn't entirely clear. It's up to the reader's expertise to understand the relationship between the numerator and denominator.

[0096] The next page gives the equation for the speed of sound = mrt. If you read the text from that page, it shows that specific heat can be calculated (scanner?), which is the specific heat ratio. If it's not an ideal gas, then there's an equation.

[0097] There's no actual text on the page specifying the "gam" and gamma of an ideal gas (must they be in the same gas?). It's not explicit, but it's obvious to anyone who knows the domain (which is why it's not explicit, because it assumes everyone knows it). But the computer knows nothing. So, we must be explicit when the computer decides how to use this equation. Similarly, "r" is the gas constant, and it must be the gas constant for "that" gas, not just any gas. So our goal is to make these things explicit.

[0098] Then, if we jump to page 5, this will show how we can use ontology to connect all of these. So we might have a simple classifier to accumulate physical things. Because this is a subclass of gas, this is a subclass of matter, and is a subclass of gas, we might have the property of the Mach number of a specific physical thing, a physical object, because only objects can have Mach number #, so it must be a physical object scaled by air. This means there is only 1 Mach number # (it is a normalized unitless thing - neither a unit nor anything else). Then part of the domain ontology shows the relevant subclasses in the domain. Gases have "molecular clusters". This is probably because you can calculate the gas constant from the general gas constant and molecular weight, you have to know more about the mass you are calculating. gamma is a constant related to gases, and so on.

[0099] What is explicitly captured is gas...etc. Gases are in the domain of gas constant and gamma, if we (11:22)..., these are properties of gases. Then we enter the enhancement type.

[0100] Enhanced types. Using this model, if you go to page 7, you can see how we can use the concept of enhanced types. Therefore, equation number 5 is for calculating the speed of sound.

[0101] Equations 1-4 lack any "enhanced type" information. This is why you find them in textbooks and the like. Now, we'll go back and add enhanced types to explicitly capture the proper use of these equations.

[0102] First, Equation 5 comes from a Java applet on NASA's website, which calculates the speed of sound.

[0103] Therefore, the input is: TGRQ. In this case, just for fun, we input the unit "us", which is the unit system specification, and it returns double.

[0104] Therefore, (T) is the temperature of the gas, expressed in Kelvin or Rankine. This tells us that it can be any of those units, but it must be an absolute temperature.

[0105] (G) is the gamma of "the" gas. They use the indefinite article just like in English. In the first article, we say "a" gas, which is a variable, meaning "a certain" gas. Then, when we say the gamma of "the" gas, it must be the same gamma that was previously identified.

[0106] (R) is the gas constant of "the" gas. If it is a metric system (consistent with Kelvin), it must be "g / mole"; if it is an imperial system, it must be "lbm / lbmole".

[0107] (Q) is theta, a constant expressed in Kelvin or Rankine. So you have two examples of this unit theta: one for the metric system and one for the imperial system. We need to find a suitable unit.

[0108] (us) is the unit system, either metric or imperial. Therefore, if it is metric, we must use the first set of units in all these enhancement types, and if it is imperial, we use the second set of units.

[0109] We use this formula to determine the speed of sound of that same gas, "the gas," in meters per second or feet per second. These equations are all independent of each other.

[0110] Equation 8 is particularly interesting. In #8 (calculating Mach), we capture this relationship at the beginning.

[0111] Therefore, the first parameter (alt) is the height of the physical object, and the physical object is moving in some kind of air. So we've connected things together.

[0112] So we say we have a physical object, this is its height, and this physical object moves in the air. From there we say "the" air and "the" physical object because we have determined how the air within the physical object is related to each other. Then we return to the Mach number and physical logic, so we use the ontology as a graph, and we use a graph mode to relate all the parameters to each other in the equation return.

[0113] Stop the competitors?

[0114] Use a graphical model to capture the relationships between inputs and outputs—that is, the relationships between all inputs and each other, and the relationships between outputs and all inputs—for any type of scientific equation.

[0115] It doesn't necessarily have to be scientific, but it's correct in any domain. In any domain, we build computational models where we compute things. We need to actually know how to choose the appropriate inputs, and we need to know what the output means in domain terminology. This is how artificial intelligence can apply the knowledge represented in the equations, the kind of approach humans would use (i.e., POSITA).

[0116] Figure 3 (P4). This is the interface of a small program from the NASA website. In this case, all the knowledge captured in the augmented type is built into the user interface. But this is not built into the way that artificial intelligence can reason about it, know what happened, and intelligently apply equations. This only involves the user, who cannot input inconsistent values ​​into the model.

[0117] Invention. The best capture lies in:

[0118] Figure 6 The combination of (ontology graphs) defines the ontology graphically. This is the ontology used to eliminate ambiguity in the meaning of these equations (5-8).

[0119] Footnote: Attribution provided. Drawings on pages 2 and 3 are from NASA.

[0120] Figure 3 shows the interface of a small program from NASA's website, where users can input numbers and use equations to calculate values... such as speed. Inputting altitude and the spacecraft's speed is the input; speed is not the output.

Claims

1. A system for the flight of an aircraft, characterized in that, include: One or more input devices, the one or more input devices being configured to input information into the system; One or more output devices; and At least one processor, operatively coupled to the one or more input devices and the one or more output devices, the at least one processor executing one or more processing modules, the one or more processing modules including an enhancement type module configured to: A graphical mode is created to capture domain knowledge of the relationship between input and output parameters of a computational expression representing flight parameters related to the flight of the aircraft, wherein the relationship between the input and output parameters includes the relationship between all input parameters and each other, as well as the relationship between the output parameter and all input parameters. Export enhanced type information from the aforementioned graphical mode; and The exported enhanced type information is added to the calculation expression of the flight parameters related to the flight of the aircraft. This addition helps the computing device to apply the knowledge represented in the calculation expression of the flight parameters related to the flight of the aircraft in the same way that a person with ordinary skills in the art can do, to calculate the flight parameters related to the flight of the aircraft.

2. The system according to claim 1, characterized in that, The graphical pattern mentioned therein includes predicate logic.

3. The system according to claim 1, characterized in that, The graphical mode mentioned therein includes concept maps.

4. The system according to claim 1, characterized in that, The graphical mode mentioned therein includes at least one of the ISO standard General Logic (CL) and Knowledge Exchange Format (KIF).

5. The system according to claim 1, characterized in that, The graphical model mentioned above was created using a web ontology language.

6. The system according to claim 1, characterized in that, The graphical pattern is formed by at least one of hierarchy, class, subclass and attribute.

7. The system according to claim 1, characterized in that, The enhanced type information mentioned therein includes the threshold model graphical pattern.

8. A tangible computer-readable medium having computer-executable instructions stored thereon, characterized in that, When executed by a processor included in a system for flight of an aircraft, the computer-executable instructions cause the processor to execute one or more processing modules, the one or more processing modules including enhancement type modules configured to: A graphical mode is created to capture domain knowledge of the relationship between input and output parameters of a computational expression representing flight parameters related to the flight of the aircraft, wherein the relationship between the input and output parameters includes the relationship between all input parameters and each other, as well as the relationship between the output parameter and all input parameters. Export enhanced type information from the aforementioned graphical mode; and The exported enhanced type information is added to the calculation expression of the flight parameters related to the flight of the aircraft. This addition helps the computing device to apply the knowledge represented in the calculation expression of the flight parameters related to the flight of the aircraft in the same way that a person with ordinary skills in the art can do, to calculate the flight parameters related to the flight of the aircraft.

9. The tangible computer-readable medium according to claim 8, characterized in that, The graphical pattern mentioned therein includes predicate logic.

10. The tangible computer-readable medium according to claim 8, characterized in that, The graphical mode mentioned therein includes concept maps.

11. The tangible computer-readable medium according to claim 8, characterized in that, The graphical mode mentioned therein includes at least one of the ISO standard General Logic (CL) and Knowledge Exchange Format (KIF).

12. The tangible computer-readable medium according to claim 8, characterized in that, The graphical model mentioned above was created using a web ontology language.

13. The tangible computer-readable medium according to claim 8, characterized in that, The graphical pattern is formed by at least one of hierarchy, class, subclass and attribute.

14. The tangible computer-readable medium according to claim 8, characterized in that, The enhanced type information mentioned therein includes the threshold model graphical pattern.

15. A method for flying an aircraft, characterized in that, The method is executed by a processor included in a system for flight of the aircraft, wherein the processor executes one or more processing modules, the one or more processing modules including an enhancement type module configured to: A graphical mode is created to capture domain knowledge of the relationship between input and output parameters of a computational expression representing flight parameters related to the flight of the aircraft, wherein the relationship between the input and output parameters includes the relationship between all input parameters and each other, as well as the relationship between the output parameter and all input parameters. Export enhanced type information from the aforementioned graphical mode; and The exported enhanced type information is added to the calculation expression of the flight parameters related to the flight of the aircraft. This addition helps the computing device to apply the knowledge represented in the calculation expression of the flight parameters related to the flight of the aircraft in the same way that a person with ordinary skills in the art can do, to calculate the flight parameters related to the flight of the aircraft.

16. The method according to claim 15, characterized in that, The enhanced type information mentioned therein includes the threshold model graphical pattern.

17. The method according to claim 15, characterized in that, The graphical model mentioned above was created using a web ontology language.

18. The method according to claim 15, characterized in that, The graphical pattern is formed by at least one of hierarchy, class, subclass and attribute.

19. The method according to claim 15, characterized in that, The graphical model mentioned above was created using a web ontology language.

20. The method according to claim 15, characterized in that, The method described therein includes at least one of the ISO standard General Logic (CL) and Knowledge Exchange Format (KIF).

Images