Perform hierarchical simplification of the learned model

Abstract

This disclosure relates to a hierarchical simplification of the execution learning model. A computer-implemented method according to one embodiment includes: applying a first input instance to a first model within a tree structure; activating a second model within the tree structure based on the first model's identification of a first topic within the first input instance; applying a second input instance to the first model and the second model; activating a third model within the tree structure based on the second model's identification of a second topic within the second input instance; applying a third input instance to the first model, the second model, and the third model; and having the third model use the third input instance to output the identification of a third topic.

Description

Technical Field

[0001] The present invention relates to machine learning, and more particularly, the present invention relates to training and utilizing neural networks. Background Art

[0002] Machine learning is often used to provide data analysis. For example, neural networks can be used to identify predetermined data within a provided input. However, these neural networks are often complex and have a large number of inputs and outputs. As a result, creating and preparing the training data required to train these neural networks is resource-intensive and time-consuming. Therefore, there is a need to simplify the organization or neural networks to simplify and reduce the amount of training data required to train such neural networks. Summary of the invention

[0003] A computer-implemented method according to one embodiment includes: applying a first input instance to a first model within a tree structure; activating a second model within the tree structure based on recognition of a first topic within the first input instance by the first model; applying a second input instance to the first model and the second model; activating a third model within the tree structure based on recognition of a second topic within the second input instance by the second model; applying a third input instance to the first model, the second model, and the third model; and outputting, by the third model, recognition of a third topic using the third input instance.

[0004] According to another embodiment, a computer program product for performing hierarchical simplification of a learning model includes a computer-readable storage medium, in which program instructions are embodied, wherein the computer-readable storage medium is not the transient signal itself, and wherein the program instructions can be executed by a processor to cause the processor to perform a method, the method comprising: applying a first input instance to a first model within a tree structure by the processor; activating a second model within the tree structure based on the recognition of a first topic within the first input instance by the processor by the first model; applying a second input instance to the first model and the second model by the processor; activating a third model within the tree structure based on the recognition of a second topic within the second input instance by the processor by the second model; applying a third input instance to the first model, the second model, and the third model by the processor; and outputting an identification of a third topic by the third model using the processor and the third input instance.

[0005] A system according to another embodiment includes a processor and logic that is integrated with the processor and can be executed by the processor, or is integrated with the processor and can be executed by the processor, wherein the logic is configured to: apply a first input instance to a first model within a tree structure; activate a second model within the tree structure based on recognition of a first topic within the first input instance by the first model; apply a second input instance to the first model and the second model; activate a third model within the tree structure based on recognition of a second topic within the second input instance by the second model; apply a third input instance to the first model, the second model, and the third model; and output recognition of a third topic by the third model using the third input instance.

[0006] A computer-implemented method according to another embodiment includes: identifying a complex model that determines multiple topics within input data; decomposing the complex model into multiple simplified models, wherein each simplified model is associated with one topic among the multiple topics and identifies the one topic among the multiple topics within the input data; determining relationships between the multiple topics; arranging the multiple simplified models into a hierarchical tree structure based on the relationships between the multiple topics; training each simplified model of the multiple simplified models within the hierarchical tree structure; and applying the trained multiple simplified models to the input data.

[0007] According to another embodiment, a computer program product for performing hierarchical simplification of a learning model includes a computer-readable storage medium, in which program instructions are embodied, wherein the computer-readable storage medium is not the transient signal itself, and wherein the program instructions can be executed by a processor to cause the processor to perform a method, the method comprising: the processor identifies a complex model that determines multiple topics within input data; the processor decomposes the complex model into multiple simplified models, wherein each simplified model is associated with one of the multiple topics and identifies the one of the multiple topics within the input data; the processor determines the relationship between the multiple topics; the processor arranges the multiple simplified models into a hierarchical tree structure based on the relationship between the multiple topics; the processor trains each of the multiple simplified models within the hierarchical tree structure; and the processor applies the trained multiple simplified models to the input data.

[0008] Other aspects and embodiments of the invention will become apparent from the following detailed description, which, when taken in conjunction with the accompanying drawings, illustrates, by way of example, the principles of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 A network architecture according to one embodiment is shown.

[0010] Figure 2According to one embodiment, Figure 1 A representative hardware environment associated with the servers and / or clients in .

[0011] Figure 3 A method for performing hierarchical simplification of a learning model according to one embodiment is shown.

[0012] Figure 4 A method for arranging neural network models in a hierarchical tree structure according to one embodiment is shown.

[0013] Figure 5 An exemplary model tree structure according to one embodiment is shown.

[0014] Figure 6 A parent / child relationship tree is shown according to one embodiment.

[0015] Figure 7 A specific application of a parent / child relationship tree to input data is shown according to one embodiment. DETAILED DESCRIPTION

[0016] The following description discloses several preferred embodiments of systems, methods, and computer program products for performing hierarchical simplification of learning models. Various embodiments provide a method for hierarchically arranging and applying a set of separate topic identification models to input data within a tree structure.

[0017] The following description is for the purpose of illustrating the general principles of the invention and is not meant to limit the inventive concepts claimed herein.In addition, the specific features described herein may be used in combination with other described features in each of the various possible combinations and permutations.

[0018] Unless expressly defined otherwise herein, all terms should be given the broadest possible interpretation, including the meaning implicit from the specification and the meaning understood by those skilled in the art and / or the meaning as defined in dictionaries, treatises, etc.

[0019] It must also be noted that, as used in the specification and the appended claims, the singular forms "a", "an", and "the" include plural referents unless otherwise specified. It will be further understood that when used in this specification, the use of the terms "comprises" and / or "comprising" specifies the presence of stated features, integers, steps, operations, elements, and / or components, but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0020] The following description discloses several preferred embodiments of systems, methods, and computer program products for performing hierarchical simplification of learning models.

[0021] In a general embodiment, a computer-implemented method includes: applying a first input instance to a first model within a tree structure; activating a second model within the tree structure based on the first model's recognition of a first topic within the first input instance; applying a second input instance to the first model and the second model; activating a third model within the tree structure based on the second model's recognition of a second topic within the second input instance; applying a third input instance to the first model, the second model, and the third model; and outputting, by the third model, an identification of a third topic using the third input instance.

[0022] In another general embodiment, a computer program product for performing hierarchical simplification of a learning model includes a computer-readable storage medium having program instructions embodied therein, wherein the computer-readable storage medium is not the transient signal itself, and wherein the program instructions are executable by a processor to cause the processor to perform a method comprising: applying a first input instance by the processor to a first model within a tree structure; activating a second model within the tree structure based on recognition by the processor of a first topic within the first input instance by the first model; applying a second input instance by the processor to the first model and the second model; activating a third model within the tree structure based on recognition by the processor of a second topic within the second input instance by the second model; applying a third input instance by the processor to the first model, the second model, and the third model; and outputting recognition of a third topic using the processor and the third input instance through the third model.

[0023] In another general embodiment, a system includes a processor and logic that is integrated with the processor, executable by the processor, or integrated with the processor and executable by the processor, wherein the logic is configured to: apply a first input instance to a first model within a tree structure; activate a second model within the tree structure based on recognition of a first topic within the first input instance by the first model; apply a second input instance to the first model and the second model; activate a third model within the tree structure based on recognition of a second topic within the second input instance by the second model; apply a third input instance to the first model, the second model, and the third model; and output, by the third model, an identification of a third topic using the third input instance.

[0024] In another general embodiment, a computer-implemented method includes: identifying a complex model that determines multiple topics within input data; decomposing the complex model into multiple simplified models, wherein each simplified model is associated with one of the multiple topics and identifies the one of the multiple topics within the input data; determining relationships between the multiple topics; arranging the multiple simplified models into a hierarchical tree structure based on the relationships between the multiple topics; training each of the multiple simplified models within the hierarchical tree structure; and applying the trained multiple simplified models to the input data.

[0025] In another general embodiment, a computer program product for performing hierarchical simplification of a learning model includes a computer-readable storage medium having program instructions embodied therein, wherein the computer-readable storage medium is not the transient signal itself, and wherein the program instructions are executable by a processor to cause the processor to perform a method comprising: identifying, by the processor, a complex model that determines multiple topics within input data; decomposing, by the processor, the complex model into multiple simplified models, wherein each simplified model is associated with one of the multiple topics and identifies the one of the multiple topics within the input data; determining, by the processor, relationships between the multiple topics; arranging, by the processor, the multiple simplified models into a hierarchical tree structure based on the relationships between the multiple topics; training, by the processor, each of the multiple simplified models within the hierarchical tree structure; and applying, by the processor, the trained multiple simplified models to the input data.

[0026] Figure 1 FIG. 1 shows an architecture 100 according to one embodiment. Figure 1 As shown, a plurality of remote networks 102 are provided, including a first remote network 104 and a second remote network 106. A gateway 101 may be coupled between the remote network 102 and the adjacent network 108. In the context of the present architecture 100, the networks 104, 106 may each take any form, including but not limited to a LAN, a WAN (such as the Internet), a public switched telephone network (PSTN), an internal telephone network, and the like.

[0027] In use, gateway 101 acts as an entry point from remote network 102 to adjacent network 108. As such, gateway 101 may act as a router capable of directing a given data packet arriving at gateway 101, as well as a switch providing the actual path in and out of gateway 101 for a given packet.

[0028] Also included is at least one data server 114 coupled to the neighboring network 108 and accessible from the remote network 102 through the gateway 101. It should be noted that the data server 114 may include any type of computing device / groupware. A plurality of user devices 116 are coupled to each data server 114. The user devices 116 may also be directly connected via one of the networks 104, 106, 108. Such user devices 116 may include a desktop computer, a laptop computer, a handheld computer, a printer, or any other type of logic. It should be noted that in one embodiment, the user device 111 may also be directly coupled to any network.

[0029] A peripheral device 120 or a series of peripheral devices 120, such as a fax machine, a printer, a networked and / or local storage unit or system, etc., may be coupled to one or more of the networks 104, 106, 108. It should be noted that the database and / or additional components may be used with or integrated into any type of network element coupled to the networks 104, 106, 108. In the context of this specification, a network element may refer to any component of a network.

[0030] According to some embodiments, the methods and systems described herein may be implemented and / or implemented on a virtual system and / or a system that emulates one or more other systems (e.g., a UNIX system that emulates an IBM z / OS environment, a UNIX system that virtually hosts a MICROSOFT WINDOWS environment, a MICROSOFT WINDOWS system that emulates an IBM z / OS environment, etc.) In some embodiments, this virtualization and / or emulation may be enhanced by using VMWARE software.

[0031] In further embodiments, one or more networks 104, 106, 108 may represent a cluster of systems commonly referred to as a "cloud." In cloud computing, shared resources such as processing power, peripherals, software, data, servers, etc. are provided to any system in the cloud on an on-demand basis, allowing services to be accessed and distributed across multiple computing systems. Cloud computing typically involves Internet connections between systems running in the cloud, but other technologies for connecting systems may also be used.

[0032] Figure 2 According to one embodiment, Figure 1 1 and 12. The diagram shows a typical hardware configuration of a workstation having a central processing unit 210 such as a microprocessor and a plurality of other units interconnected by a system bus 212.

[0033] Figure 2The workstation shown includes random access memory (RAM) 214, read only memory (ROM) 216, an I / O adapter 218 for connecting peripheral devices such as a disk storage unit 220 to the bus 212, a user interface adapter 222 for connecting a keyboard 224, a mouse 226, a speaker 228, a microphone 232 and / or other user interface devices (such as a touch screen and a digital camera (not shown)) to the bus 212, a communications adapter 234 for connecting the workstation to a communications network 235 (e.g., a data processing network), and a display adapter 236 for connecting the bus 212 to a display device 238.

[0034] A workstation may have an operating system resident on it, such as Microsoft Operating system (OS), MACOS, UNIX OS, etc. It should be understood that the preferred embodiment can also be implemented on platforms and operating systems other than the above. XML, C and / or C++ language or other programming languages ​​and object-oriented programming methods can be used to write the preferred embodiment. Object-oriented programming (OOP) can be used, which has been increasingly used to develop complex applications.

[0035] Reference now Figure 3 , shows a flow chart of a method 300 according to an embodiment. The method 300 may be performed according to the present invention. Figure 1-2 and 5-6, especially in various embodiments. Of course, as those skilled in the art will understand when reading this specification, Figure 3 More or fewer of the operations specifically described in may be included in method 300 .

[0036] Each step of method 300 may be performed by any suitable component of an operating environment. For example, in various embodiments, method 300 may be performed in part or in whole by one or more servers, computers, or some other device having one or more processors. A processor (e.g., a processing circuit, chip, and / or module implemented in hardware and / or software, and preferably having at least one hardware component) may be utilized in any device to perform one or more steps of method 300. Illustrative processors include, but are not limited to, a central processing unit (CPU), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), and the like, combinations thereof, or any other suitable computing device known in the art.

[0037] like Figure 3As shown, method 300 may begin with operation 302, where a first input instance is applied to a first model within a tree structure. In one embodiment, the first model may include a learning model, such as a first neural network. In another embodiment, the tree structure may represent multiple individual models and relationships between the models. For example, each model within the tree structure may include a learning model, such as a neural network.

[0038] In addition, in one embodiment, the tree structure may include a root model, one or more intermediate models, and one or more terminal models. For example, the root model may include an initial model on which all other models in the tree structure depend. For example, the root model may not depend on any other model in the tree structure. In another example, the intermediate model may include a model in the tree structure that depends on another model but also has a model that depends on it (e.g., a child model in the tree structure, etc.). In yet another example, the terminal model may include a model that depends on another model but no model depends on it (e.g., a leaf model in the tree structure, etc.).

[0039] In addition, in one embodiment, the tree structure can be arranged based on topics. For example, each of the multiple models can be associated with a single topic that is different from other models. For example, each of the multiple models can store a sequence of words for a separate topic. The topics can include keywords, variations of keywords, etc. In another example, each of the multiple models can analyze the input to determine whether a single topic associated with the model is found in the input. In yet another example, each model can be labeled with a single topic associated with it.

[0040] Additionally, in one embodiment, each of the multiple topics may be analyzed to determine relationships between the topics. In another embodiment, parent / child topics may be determined within the multiple topics. For example, within the input provided, a first topic may always be found to precede a second topic. In another example, the first topic may then be identified as a parent of the second topic, and the second topic may be identified as a child of the first topic.

[0041] Likewise, in one embodiment, multiple models may be arranged in a tree structure based on these topics / relationships. For example, a child model may be arranged as a child of a parent model in a tree structure. In the above example, the second topic may be arranged as a child of the first topic in a tree structure.

[0042] Additionally, in one embodiment, the first model may include a root model within a tree structure. In another embodiment, the first model may include a classification module that outputs a label (e.g., a topic) based on the provided input. For example, the label may include an identification of a predetermined topic within the provided input.

[0043] In addition, in one embodiment, the first input instance may include text data, audio data, time series data, etc. In another embodiment, the first input instance may include a first portion of the input data. For example, the input data may include a text document, an audio recording, etc. In another example, the input data may be divided into multiple portions. In yet another example, the multiple portions may be arranged in chronological order (e.g., such that the first portion is located before the second portion, the second portion is located before the third portion, and so on).

[0044] Additionally, method 300 may continue with operation 304, where a second model within the tree structure is activated based on the first model's identification of a first topic within the first input instance. In one embodiment, the first input instance may be analyzed by the first model, wherein the first model is associated with the first topic. In another embodiment, the first model may identify the first topic within the first input instance.

[0045] Likewise, in one embodiment, in response to the identification of the first topic within the first input instance, all descendants of the first model within the tree structure may be activated. For example, the second model may include descendant models of the first model within the tree structure. In another example, the second model may be applied to subsequent inputs together with the first model.

[0046] Additionally, in one embodiment, the second model may include a learning model, such as a second neural network separate from the first neural network. In another embodiment, the second model may include an intermediate model within a tree structure. For example, the second model may have one or more children within the tree structure. In another example, the second model may include a classification module that outputs a label (e.g., a topic) based on the input provided.

[0047] Additionally, method 300 may continue with operation 306, where a second input instance is applied to the first model and the second model. In one embodiment, the second input instance (e.g., within a plurality of portions of the input arranged in a time sequence, etc.) may include a second portion of the input data that occurs after the first portion of the input data.

[0048] Additionally, method 300 may proceed to operation 308, where a third model within the tree structure is activated based on the second model's identification of a second topic within the second input instance. In one embodiment, the second input instance may be analyzed by a first model and a second model, wherein the first model is associated with the first topic and the second model is associated with the second topic. In another embodiment, the second model may identify the second topic within the second input instance.

[0049] Likewise, in one embodiment, in response to the identification of the second topic within the first input instance, all descendants of the second model within the tree structure may be activated. For example, the third model may include descendant models of the second model within the tree structure. In another example, the third model may be applied to subsequent inputs together with the first model and the second model.

[0050] Additionally, in one embodiment, the third model may include a learning model, such as a third neural network separate from the first neural network and the second neural network. In another embodiment, the third model may include a terminal model within the tree structure. For example, the third model may have no children within the tree structure. In another example, the third model may include a classification module that outputs a label (e.g., a topic) based on the input provided.

[0051] Additionally, method 300 may continue with operation 310, where a third input instance is applied to the first model, the second model, and the third model. In one embodiment, the third input instance (e.g., within a plurality of portions of the input arranged in a time sequence, etc.) may include a third portion of the input data that occurs after the second portion of the input data.

[0052] Additionally, method 300 may proceed to operation 312, where the third model outputs an identification of a third topic using the third input instance. In one embodiment, the third model may analyze the third input instance and may output a label.

[0053] In this manner, a set of separate topic identification models, hierarchically arranged in a tree structure by the topics they identify, can be applied to input data, wherein models within the set are activated and applied to later portions of the input data based on earlier topic identifications in earlier portions of the input data. This set of separate models can have much less complexity than a single model that performs identification of multiple topics, and therefore, can require much less training data and training time than a single model. This can reduce the amount of storage space, processor utilization, and memory usage used to train and implement topic identification models, which can improve the performance of computing devices performing such training and implementation.

[0054] Reference now Figure 4 , shows a flow chart of a method 400 for arranging a neural network model in a hierarchical tree structure according to one embodiment. The method 400 may be used in the present invention. Figure 1-2 and 5-6, especially in any environment described in the various embodiments. Of course, as those skilled in the art will understand when reading this specification, Figure 4 More or fewer of the operations specifically described may be included in method 400 .

[0055] Each step of method 400 may be performed by any suitable component of an operating environment. For example, in various embodiments, method 400 may be performed in part or in whole by one or more servers, computers, or some other device having one or more processors therein. A processor (e.g., a processing circuit, chip, and / or module implemented in hardware and / or software, and preferably having at least one hardware component) may be utilized in any device to perform one or more steps of method 400. Illustrative processors include, but are not limited to, a central processing unit (CPU), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), and the like, combinations thereof, or any other suitable computing device known in the art.

[0056] like Figure 4 As shown, method 400 may begin at operation 402, where a complex model that determines multiple topics within input data is identified. In one embodiment, the complex model may include a single neural network.

[0057] Additionally, method 400 may proceed to operation 404, in which the complex model is decomposed into a plurality of simplified models, wherein each simplified model is associated with a topic of a plurality of topics, and the one of the plurality of topics within the input data is identified. In one embodiment, each simplified model may be associated with a topic that is different from other topics associated with other models (e.g., each topic may be unique).

[0058] Further, method 400 may proceed to operation 406, where a relationship between the plurality of topics is determined. In one embodiment, the relationship may be predefined, may be determined based on topic relationship analysis, etc. In another embodiment, parent / child topics may be determined among the plurality of topics.

[0059] Furthermore, method 400 may proceed to operation 408, where the plurality of simplified models are arranged into a hierarchical tree structure based on the relationships among the plurality of topics. In one embodiment, the child models of a given parent model may be arranged as children of the parent model within the tree structure.

[0060] Additionally, the method 400 may proceed to operation 410, in which each of the plurality of simplified models within the hierarchical tree structure is trained. In one embodiment, each of the plurality of simplified models may be trained using predetermined instances of training data.

[0061] Additionally, method 400 may proceed to operation 412, in which the trained plurality of simplified models are applied to the input data. In one embodiment, the input data may include sequentially organized data. For example, the input data may have a consistent order, wherein a first portion of the input data always appears before a second portion of the input data. In another embodiment, a predetermined simplified model (e.g., a root model or a direct descendant of the root model, etc.) may be initially applied to a first portion of the input data. For example, the first portion of the input data may include a predetermined portion of the input data within a sequential organization.

[0062] In addition, in one embodiment, in response to the identification of the subject by the predetermined simplified model, a child model of the predetermined simplified model within the tree structure can be activated and applied to the second portion of the input data. In another embodiment, model activation can be performed until the input data is completely processed, or the terminal model of the predetermined simplified model is activated and applied.

[0063] As a result, the amount of training data required to train multiple simplified models may be less than the amount of training data required to train a complex model. For example, if a complex model has M inputs and N outputs, then training data on the order of M×N is required to train the complex model. By decomposing the complex model into M simplified models, each with one input, training data on the order of M+N is required to train the simplified models. This reduces the amount of training data required during subject recognition, and may reduce storage, processing, and resource utilization of a computing device performing such training, which may improve the performance of such a computing device.

[0064] Figure 5 An exemplary model tree structure 500 according to an exemplary embodiment is shown. As shown, a plurality of models 502-514 are arranged in the tree structure 500. In one embodiment, each of the plurality of models 502-514 may include a single independent neural network. In another embodiment, each of the plurality of models 502-514 may take text, audio, and / or time series data as input, and may search for a predetermined topic within the input. For example, each of the plurality of models 502-514 may search for a predetermined topic different from the other plurality of models 502-514, and if a predetermined topic is identified in the input, a first predetermined value may be output (if a predetermined topic is not identified in the input, a second predetermined value may be output).

[0065] Additionally, in one embodiment, each of the plurality of models 502-514 may be associated with a predetermined topic, and the arrangement of the tree structure 500 may be based on the relationships between the topics. For example, each of the plurality of models 502-514 may be associated with a predetermined topic, wherein the predetermined topic includes the topic searched by the model. Predetermined parent / child relationships between each topic may be provided, and these relationships may be used to create the tree structure 500.

[0066] For example, the provided parent / child relationship may indicate that the subject searched by the second model 504 and the subject searched by the third model 506 are subordinate to the subject searched by the first model 502, and as a result, the second model 504 and the third model 506 are arranged within the tree structure 500 as children of the first model 502. Similarly, the provided parent / child relationship may indicate that the subject searched by the fourth model 508 and the subject searched by the fifth model 510 are subordinate to the subject searched by the second model 504, and as a result, the fourth model 508 and the fifth model 510 are arranged within the tree structure 500 as children of the second model 504. Furthermore, the provided parent / child relationship may indicate that the subject searched by the sixth model 512 and the subject searched by the seventh model 514 are subordinate to the subject searched by the third model 506, and as a result, the sixth model 512 and the seventh model 514 are arranged within the tree structure 500 as children of the third model 506.

[0067] In addition, it may be determined that the fourth model 508, the fifth model 510, the sixth model 512, and the seventh model 514 do not have any child models. As a result, these models may be arranged as terminal nodes within the tree structure 500. Since the second model 504 and the third model 506 have child nodes, these models may be arranged as intermediate nodes within the tree structure 500.

[0068] In addition, in one embodiment, only the first model 502 subordinate to the root 516 of the tree structure 500 may be activated and provided with the first input instance. In another embodiment, the first input instance may include a first portion of a plurality of sequentially organized input instances. In yet another embodiment, the first model 502 may be associated with a first predetermined topic, and the first predetermined topic within the first input instance may be searched.

[0069] In addition, in response to the first model 502 identifying the first predetermined topic within the first input instance, all descendants of the first model 502 within the tree structure 500 (e.g., the second model 504 and the third model 506) are activated and provided with the second input instance together with the first model 502. In one embodiment, the second input instance may include a second portion of a plurality of sequentially organized input instances that appear immediately after the first input instance. In another embodiment, the second model 504 and the third model 506 may be associated with the second predetermined topic and the third predetermined topic, respectively, and their predetermined topics within the second input instance may be searched.

[0070] In addition, in response to the second model 504 identifying the second predetermined topic within the second input instance, all descendants of the second model 504 in the tree structure 500 (e.g., the fourth model 508 and the fifth model 510) are activated and provided with the third input instance together with the first model 502, the second model 504, and the third model 506. In one embodiment, the third input instance may include a third portion of a plurality of sequentially organized input instances that appears immediately after the second input instance. In another embodiment, the fourth model 508 and the fifth model 510 may be associated with the fourth predetermined topic and the fifth predetermined topic, respectively, and their predetermined topics within the third input instance may be searched.

[0071] In this manner, each of the plurality of models 502-514 may be trained to identify a single associated topic within the input, rather than training a single model to identify all associated topics. This may reduce the amount of resources utilized by a computing device performing the training, thereby improving the performance of the computing device. Additionally, the plurality of models 502-514 may be selectively applied to the input according to their arrangement within the tree structure 500, and thus associated topics may be identified in a manner similar to a single model trained to identify all associated topics.

[0072] Generate machine-readable business process definitions from text documents written in natural human languages

[0073] In natural language processing using machine learning, the main problem of accuracy enhancement is to ensure sufficient learning data volume. Generally, using a learning model with high writing performance requires a learning data volume proportional to the learning performance to train the learning model. Generally, the learning volume can be considered as the number of parameters in the model.

[0074] In the case of three inputs and two outputs, the number of internal parameters (weight coefficients of the inputs) is six. If one input is allowed as a fourth input, the number of internal parameters increases to eight. Furthermore, if three outputs are provided, the number of parameters becomes twelve. In order to determine these parameters by learning, at least the same number of learning data as the parameters is required.

[0075] In order to train a model that performs efficient learning (i.e., solves a problem using the least amount of learning data), it is necessary to build a learning model that is of the smallest possible size and in a form that is optimized for the problem.

[0076] Therefore, the following approach will be adopted:

[0077] 1. Combining multiple small models to build a model with externally specified parent / child relationships.

[0078] 2. According to the parent / child relationship, based on the detection results of the data provided in the upper layer in chronological order, the constructed model operates by changing the model activated in the lower layer, while changing dynamically as a whole.

[0079] Although learning data covering all cases is required in order to learn all conditions by a single large network, combining multiple small networks enables the cost of such learning data to be reduced. In order to combine small networks, information for specifying how to combine them is required, but, here, it is assumed that such information can be defined in advance. In general, this therefore saves the learning cost of information for externally specified logical structures.

[0080] In the case of analyzing a text document written in a natural language, an overall learning model is created by creating multiple small models that store word sequences for separate topics and externally specifying parent / child relationships between topics, rather than creating one learning model that stores the work sequence included in the entire text document. This corresponds to the way people think when reading a document, that is, specifying an area as a topic to narrow down the range of topics that may be discussed later, thereby facilitating understanding because it reduces the items to be determined.

[0081] Figure 6 A parent / child relationship tree 600 is shown according to an exemplary embodiment. Each model 602-614 indicates a subject, and a directed link indicates a parent / child relationship. Within the tree 600, children are subordinate to corresponding parents.

[0082] The analysis of the text is performed sequentially from the beginning of the text based on individual sentences or parts similar to sentences. Together with the detection of the parent topic in the sentence in the analysis, the models are automatically activated, each model detecting the subtopics subordinate to the parent topic. Figure 7 A specific application 700 of the tree 600 to input data is shown in FIG.

[0083] like Figure 7 As shown, in the initial state, only the overlay model 602 is activated, and the overlay model 602 is only subordinate to Figure 6616 of the tree 600 in . When the coverage model 602 detects the topic "coverage" in the first input row 702A (e.g., in response to detecting the term "coverage"), the Figure 6 The normal_case model 604 and the exclusion model 606 of the coverage model 602 of the tree 600 in are activated and applied to the second input row 702B.

[0084] When the normal_condition model 604 detects the subject "normal condition" in the second input row 702B (e.g., in response to detecting the term "condition"), the injury model 608 and the illness model 610 are based on Figure 6 600 in the tree 600 and is applied to the third input row 702C. In the third input row 702C, the injury model 608 detects the subject "injury" (e.g., in response to detecting the term "injury"), and the illness model 610 detects the subject "illness" (e.g., in response to detecting the term "illness").

[0085] When the exclusion model 606 detects the topic "exclusion" in the fourth input row 702D (e.g., in response to detecting the term "exclusion"), the injury model 608 and the illness model 610 are deactivated and instead, the Figure 6 Exemption 1 model 612 and exemption 2 model 614 of exclusion model 606 within tree 600 are activated.

[0086] In fifth input row 702E and sixth input row 702F, respectively, exemption 1 model 612 detects the topic "Exemption 1" (e.g., in response to detecting the term "first exemption"), and exemption 2 model 614 detects the topic "Exemption 2" (e.g., in response to detecting the term "second exemption").

[0087] In one embodiment, based on the indication of the parent / child relationship, the analysis engine can operate while changing the activated model. Table 1 shows an example code to implement such an analysis engine.

[0088] Table 1

[0089]

[0090]

[0091] In the case of combining a component with M options and a component with N options, there are M×N options. In order to set this model by machine learning, learning data of the order of M×N is required. On the other hand, if the components are learned separately, only learning data of the order of M+N is required. By explicitly indicating that two components are logically independent in the model in the form of tree separation, this can be regarded as the effect of eliminating options in unnecessary combinations.

[0092] In one embodiment, a method for marking input data includes: creating a tree-structured learning model for marking input data, including creating a tree-structured learning model for marking input data, wherein the model of the tree structure is created from an end model based on a dependency relationship. In addition, the input data organized in chronological order is read from the beginning of the input data, and the model is applied starting from the root of the tree structure. In addition, based on the detection results of the model, the model is selectively activated and applied within the tree structure. In addition, the input data is marked based on the detection results of the activated and applied model.

[0093] The present invention may be a system, a method and / or a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for causing a processor to implement various aspects of the present invention.

[0094] A computer-readable storage medium may be a tangible device that can hold and store instructions used by an instruction execution device. A computer-readable storage medium may be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples of computer-readable storage media (a non-exhaustive list) include: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a static random access memory (SRAM), a portable compact disk read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanical encoding device, such as a punch card or a raised structure in a groove on which instructions are stored, and any suitable combination of the foregoing. As used herein, a computer-readable storage medium is not to be interpreted as a transient signal per se, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagating through a waveguide or other transmission medium (e.g., a light pulse through a fiber optic cable), or an electrical signal transmitted through a wire.

[0095] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to each computing / processing device, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network can include copper transmission cables, optical fiber transmissions, wireless transmissions, routers, firewalls, switches, gateway computers, and / or edge servers. The network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in the computer-readable storage medium in each computing / processing device.

[0096] The computer program instructions for performing the operation of the present invention may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-related instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages, such as Smalltalk, C++, etc., and conventional procedural programming languages, such as "C" language or similar programming languages. Computer-readable program instructions may be executed entirely on a user's computer, partially on a user's computer, as an independent software package, partially on a user's computer, partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., using an Internet service provider to connect via the Internet). In some embodiments, an electronic circuit, such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), may be personalized by utilizing the state information of the computer-readable program instructions, and the electronic circuit may execute the computer-readable program instructions, thereby realizing various aspects of the present invention.

[0097] Various aspects of the present invention are described herein with reference to the flow charts and / or block diagrams of the methods, devices (systems) and computer program products according to embodiments of the present invention. It should be understood that each box of the flow chart and / or block diagram and the combination of each box in the flow chart and / or block diagram can be implemented by computer-readable program instructions.

[0098] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, thereby producing a machine, so that when these instructions are executed by the processor of the computer or other programmable data processing device, a device that implements the functions / actions specified in one or more boxes in the flowchart and / or block diagram is generated. These computer-readable program instructions can also be stored in a computer-readable storage medium, and these instructions cause the computer, programmable data processing device, and / or other equipment to work in a specific manner, so that the computer-readable medium storing the instructions includes a manufactured product, which includes instructions for implementing various aspects of the functions / actions specified in one or more boxes in the flowchart and / or block diagram.

[0099] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device so that a series of operating steps are performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to implement the functions / actions specified in one or more boxes in the flowchart and / or block diagram.

[0100] The flow chart and block diagram in the accompanying drawings show the possible architecture, function and operation of the system, method and computer program product according to multiple embodiments of the present invention. In this regard, each square box in the flow chart or block diagram can represent a part of a module, program segment or instruction, and a part of the module, program segment or instruction includes one or more executable instructions for realizing the specified logical function. In some alternative implementations, the function marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two continuous square boxes can actually be executed substantially in parallel, and they can sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs the specified function or action, or can be implemented with a combination of special hardware and computer instructions.

[0101] In addition, a system according to various embodiments may include a processor and logic integrated with the processor and / or executable by the processor, the logic being configured to perform one or more processing steps described herein. By integrated with, it is meant that the processor has logic embedded therewith as hardware logic, such as an application specific integrated circuit (ASIC), FPGA, and the like. By executable by a processor, it is meant that the logic is hardware logic; software logic, such as firmware, part of an operating system, part of an application; and the like, or some combination of hardware and software logic that is accessible to the processor and configured to cause the processor to perform certain functions when executed by the processor. As is known in the art, software logic may be stored on local and / or remote memory of any memory type. Any processor known in the art may be used, such as a software processor module and / or a hardware processor, such as an ASIC, FPGA, a central processing unit (CPU), an integrated circuit (IC), a graphics processing unit (GPU), and the like.

[0102] It will be clear that the various features of the aforementioned systems and / or methods may be combined in any manner, thereby creating multiple combinations from the description presented above.

[0103] It will be further appreciated that embodiments of the present invention may be provided as a service deployed on behalf of a client to provide services on demand.

[0104] Although various embodiments have been described above, it should be understood that they are exemplary only and not restrictive.Thus, the breadth and scope of a preferred embodiment should not be limited by any of the above-described exemplary embodiments, but should be defined only in accordance with the following claims and their equivalents.

Claims

1. A computer-implemented method comprising: Applying the first input instance only to the first model within the tree structure; activating a second model within the tree structure based on identification by the first model of a first topic within the first input instance; applying a second input instance different from the first input instance to the first model and the second model; activating a third model within the tree structure based on identification by the second model of a second topic within the second input instance; applying a third input instance different from the first input instance and the second input instance to the first model, the second model, and the third model; and outputting, by the third model, an identification of a third subject using the third input instance, Wherein, the first input instance, the second input instance, and the third input instance come from a single input data source organized sequentially, wherein the first input instance is located before the second input instance in the single input data source, and wherein the second input instance is located before the third input instance in the single input data source.

2. The computer-implemented method of claim 1 , wherein: The first model includes a first neural network.

3. The computer-implemented method of claim 1 , wherein: The tree structure represents a plurality of individual models and the relationships between the individual models.

4. The computer-implemented method of claim 1 , wherein: The tree structure is arranged based on topics.

5. The computer-implemented method of claim 1 , wherein: The first model includes a classification module that outputs a topic based on the input provided.

6. The computer-implemented method of claim 1 , wherein: The first input instance is selected from the group consisting of text data, audio data, and time series data.

7. The computer-implemented method of claim 1 , wherein: In response to the identification of the first topic within the first input instance, all children of the first model within the tree structure are activated.

8. The computer-implemented method of claim 1 , wherein: The first model includes a root model in the tree structure, the second model includes an intermediate model in the tree structure, and the third model includes a terminal model in the tree structure.

9. The computer-implemented method of claim 1 , wherein: The first input instance includes a first portion of input data, wherein the input data is divided into a plurality of time-sequentially arranged portions.

10. A computer program product for performing hierarchical simplification of a learning model, wherein: The computer program product comprises program instructions executable by a processor to cause the processor to perform the method according to any one of claims 1-9.

11. A system comprising: processor; as well as Logic, the logic is integrated with the processor and can be executed by the processor, or is integrated with the processor and can be executed by the processor, the logic being configured to perform the method according to any one of claims 1-9.

12. A system comprising modules for performing the steps of the method according to any one of claims 1-9.

13. A computer-readable storage medium for performing hierarchical simplification of a learning model, wherein program instructions are embodied in the computer-readable storage medium, wherein: The computer-readable storage medium is not a transient signal itself, and the program instructions can be executed by a processor to enable the processor to perform the method according to any one of claims 1-9.

Images