Summary of the invention
In order at least solve to train semantic resources speed slower in the prior art, recognition speed caused by flaky process is slow
Slowly, the high problem of resources occupation rate in identification.
In a first aspect, the embodiment of the present invention provides a kind of semantic resources training method for voice dialogue platform, including:
In response to the click of the publication button of speech production in voice dialogue platform, in the history training of the speech production
The first version number that training script is obtained in information, obtains the second edition number of the training script of the voice dialogue platform;
When the first version number and the second edition number are consistent, unmodified data pair in the speech production is determined
The first semantic resources answered, by corresponding second semantic resources of modification data;
The training classification of second semantic resources is determined by modification data according to described, is second semantic resources
Determine at least one son training;
By at least one described son training, third semantic resources are determined, by first semantic resources and the third
Semantic resources combine, to complete the training of the semantic resources.
Second aspect, the embodiment of the present invention provide a kind of semantic resources training system for voice dialogue platform, including:
Version number determines program module, in response in voice dialogue platform speech production publication button click,
The first version number that training script is obtained in the history training information of the speech production, obtains the voice dialogue platform
The second edition number of training script;
Semantic resources determine program module, for determining when the first version number and the second edition number are consistent
Corresponding first semantic resources of unmodified data in the speech production, by corresponding second semantic resources of modification data;
Training classification determines program module, for, by modification data, determining the instruction of second semantic resources according to described
Practice classification, determines at least one son training for second semantic resources;
Semantic resources training program module, for third semantic resources being determined, by institute by least one described son training
The first semantic resources are stated in conjunction with the third semantic resources, to complete the training of the semantic resources.
The third aspect provides a kind of electronic equipment comprising:At least one processor, and with described at least one
Manage the memory of device communication connection, wherein the memory is stored with the instruction that can be executed by least one described processor, institute
It states instruction to be executed by least one described processor, so that at least one described processor is able to carry out any embodiment of the present invention
The semantic resources training method for voice dialogue platform the step of.
Fourth aspect, the embodiment of the present invention provide a kind of storage medium, are stored thereon with computer program, and feature exists
In the semantic resources training side for voice dialogue platform of realization any embodiment of the present invention when the program is executed by processor
The step of method.
The beneficial effect of the embodiment of the present invention is:Pass through the publication button in response to speech production in voice dialogue platform
Click, can be after developer has modified in product technical ability and/or the data of dictionary, real-time update semantic resources are convenient
The debugging of developer promotes the usage experience of developer.Since which data had modified by comparison, when training, only
The part resource for needing training to modify, saves the resource of server, still can be with so that in the case where server resource is in short supply
Holding is efficiently completed semantic resources training, further increases the usage experience of developer.
Specific embodiment
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is
A part of the embodiment of the present invention, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art
Every other embodiment obtained without creative efforts, shall fall within the protection scope of the present invention.
A kind of semantic resources training method for voice dialogue platform provided as shown in Figure 1 for one embodiment of the invention
Flow chart, include the following steps:
S11:In response to the click of the publication button of speech production in voice dialogue platform, in the history of the speech production
The first version number that training script is obtained in training information, obtains the second edition of the training script of the voice dialogue platform
Number;
S12:When the first version number and the second edition number are consistent, unmodified number in the speech production is determined
According to corresponding first semantic resources, by corresponding second semantic resources of modification data;
S13:The training classification of second semantic resources is determined by modification data according to described, it is semantic for described second
Resource determines at least one son training;
S14:By it is described at least one son training, determine third semantic resources, by first semantic resources with it is described
Third semantic resources combine, to complete the training of the semantic resources.
In the present embodiment, the voice dialogue platform is the Intelligent dialogue open platform towards developer.In order to allow
Entire Intelligent dialogue content is more easier to be understood by developer, more humanized, reduces the difficulty of starting with of developer, and improves exploitation
The development rate of person.Multilayer high abstraction has been done to basic technologies such as speech recognition/semantics recognitions, be abstracted into product/technical ability/appoint
Each levels such as business/intention/saying.The semantic training of the voice dialogue platform is skill level, includes task in technical ability,
It is again intentional under task, it is intended that under be only the entity of saying and be added to word in voice technical ability to guarantee widely to identify
Library.
The semanteme training and identification of the voice dialogue platform are skill levels.Technical ability is similar to an APP, passes through language
One or more specific function is completed in sound dialogue, as wechat supports multiple functions:Message, circle of friends, payment etc., task are exactly
One of function, such as inquires address, and periphery or positioning are searched in navigation.Task is one of important composition ingredient of technical ability, is
Complete the intersection of single-wheel dialogue or the more wheel dialogues of a certain item function.The each round of user is talked with, and is construed as one
It is intended to, a task is made of one or more intention.
The function of semantics recognition is exactly to identify whether user's word has hit this technical ability, if having hit the skill
Can, that is which task and intention hit again.After semantics recognition to specific be intended to, subsequent data processing could be continued
It is shown with result.
Developer develops speech production in the voice dialogue platform, when developer is by the language in speech production
After sound technical ability and/or dictionary are developed, publication test is carried out.
For step S11, voice dialogue platform response is in the click of speech production publication button, to the speech production
History training information is inquired, and the version number of the last training script of the speech production is inquired, for example, the voice
The version number of the training script of the last time training speech production recorded in the history training information of product is 2.0;Obtain institute
The second edition number of the training script of predicate sound dialogue platform.
For step S12, such as the version number of training script of the voice dialogue platform is 2.0, the first version
It is number consistent with the second edition number, it determines in the speech production, corresponding first semantic resources of unmodified data are modified
Corresponding second semantic resources of data.Wherein, since the speech production is not to submit for the first time, for example, can be in voice pair
Old version is found in words platform, may thereby determine that out which data modification mistake in current publication speech production, which does not have
It modified, so that it is determined that having gone out corresponding first semantic resources of unmodified data in speech production, by modification data corresponding
Two semantic resources.
The training classification of second semantic resources is determined according to the data modified for step S13, for institute
It states the second semantic resources and determines at least one son training, for example, the voice technical ability in the speech production is modified:Addition and/
Or modification voice technical ability, modify the intention and/or saying in voice technical ability;Dictionary in the speech production is modified.For
In these by modification data, to determine the training classification of second semantic resources, so that it is determined that going out at least one son training.
For step S14, by least one the son training determined in step s 13, the third after determining training is semantic
Resource combines first semantic resources and the third semantic resources, and the training result of subtask and last time are instructed
The semantic resources not changed in white silk, according to semantic analysis service need format integrate, and according to training use training script/
Training time/training path generates this training information, ultimately generates semantic resources, completes semantic in the dialogue platform of voice
The training of resource.
It can be seen that by the embodiment through the point in response to the publication button of speech production in voice dialogue platform
It hits, can be after developer have modified in product technical ability and/or the data of dictionary, real-time update semantic resources facilitate exploitation
The debugging of person promotes the usage experience of developer.Since which data had modified by comparison, when training, it is only necessary to
Training sub-fraction resource, saves the resource of server, so that can still keep high in the case where server resource is in short supply
The completion semantic resources training of effect, further increases the usage experience of developer.
As an implementation, in the present embodiment, the semantic resources training classification includes:Classification of task, intention
Identification, dictionary identification;
It is described to include by modification data:Voice technical ability, intention and/or saying, dictionary;
When it is described by modification data include at least voice technical ability when, second semantic resources son training include at least appoint
Business classification based training,
When described included at least by modification data is intended to and/or when saying, the son training of second semantic resources is at least
It is trained including classification of task training and intention assessment,
When it is described by modification data include at least dictionary when, second semantic resources son training include at least dictionary know
Xun Lian not.
In the present embodiment, for example, developer is added in the weather technical ability in the voice technical ability of speech production looks into
The intention of weather is ask, also, joined the place name of some relatively unexpected winners in dictionary.The function of the speech production is expanded in this way
Can, and the speech production area is applicable in more extensively, in this way, the data modified include intention and dictionary.
When it is described by modification data include being intended to and when dictionary, the son training of second semantic resources, which includes at least, appoints
Business classification based training, intention assessment training and dictionary recognition training.
It can be seen that by the embodiment by carrying out task fractionation to the data modified, to generic task
Unified training is carried out, different tasks is directed to and carries out corresponding training, improve training effect.
As an implementation, in the present embodiment, when the first version number and the second edition number are inconsistent
When, it according to data all in the speech production, determines the training classification of the semantic resources of the speech production, is the semanteme
Resource determines at least one son training, with the training semantic resources.
In the present embodiment, when having found that trained version is inconsistent, it should in the voice dialogue platform again
Training institute is skilled.The speech recognition resources trained every time are only trained if it were not for a part by the training script of 1.0 versions, and one
Part is by the training script training of 2.0 versions, then this semantic resources is just available, voice dialogue platform can be according to its version
Called corresponding kernel version to use originally.Confirm version, be in order to guarantee that the training script of a resource is consistent, if
Inconsistent, the kernel of voice dialogue platform, which will be unable to determine, corresponds to the voice dialogue product is which version of the calling on earth
Kernel.
It can be seen that the version number by comparing training script by the embodiment, ensure that voice dialogue platform
Kernel can be explicitly called, the version disunity of the training script because of semantic resources each in same speech production is prevented, and is produced
Raw kernel calls problem.
As an implementation, in the present embodiment, the method also includes:By association Cheng Chi to it is described at least one
Son training parallel training.
General thread pool and process pool preferably handle same or similar task, compare if the input and output of task have
Big difference, then after process pool and thread pool will be continued to execute in the processing returned the result to task using call back function
The task in face increases development difficulty, is also easy to cause bug.Since there are many types for the subtask, if every kind of task is all
Increase a thread pool, then parallel effect can be far short of what is expected.If the task of polymorphic type is put into a thread pool, can increase
Increase the difficulty of hair.Therefore it employs techniques to realize an association Cheng Chi, use when only needs to add a modifier can
Training mission is added to association Cheng Chizhong to realize.Size of code is simplified, while also ensuring parallel effect.
It can be seen that by the implementation method and further accelerate trained speed by parallel training, reduce out
The waiting time of originator improves the usage experience of developer.
A kind of semantic resources training system for voice dialogue platform of one embodiment of the invention offer is provided
Structural schematic diagram, the technical solution of the present embodiment is applicable to the semantic resources training to equipment for voice dialogue platform
Method, which can be performed the semantic resources training method that voice dialogue platform is used for described in above-mentioned any embodiment, and match
It sets in the terminal.
A kind of semantic resources training system for voice dialogue platform provided in this embodiment includes:Version number determines journey
Sequence module 11, semantic resources determine that program module 12, training classification determine program module 13 and semantic resources training program module
14。
Wherein, version number determines program module 11 in response to the publication button of speech production in voice dialogue platform
It clicks, the first version number of training script is obtained in the history training information of the speech production, obtains the voice dialogue
The second edition number of the training script of platform;Semantic resources determine program module 12 for when the first version number and described the
When two version numbers are consistent, determine corresponding first semantic resources of unmodified data in the speech production, corresponded to by modification data
The second semantic resources;Training classification determines program module 13 for, by modification data, determining that described second is semantic according to described
The training classification of resource determines at least one son training for second semantic resources;Semantic resources training program module 14 is used
In by least one described son training, determines third semantic resources, first semantic resources and the third semanteme are provided
Source combines, to complete the training of the semantic resources.
Further, the semantic resources training classification includes:Classification of task, intention assessment, dictionary identification;
It is described to include by modification data:Voice technical ability, intention and/or saying, dictionary;
When it is described by modification data include at least voice technical ability when, second semantic resources son training include at least appoint
Business classification based training,
When described included at least by modification data is intended to and/or when saying, the son training of second semantic resources is at least
It is trained including classification of task training and intention assessment,
When it is described by modification data include at least dictionary when, second semantic resources son training include at least dictionary know
Xun Lian not.
Further, the system is also used to:When the first version number and the second edition number are inconsistent, according to
All data in the speech production determine the training classification of the semantic resources of the speech production, are that the semantic resources are true
At least one fixed son training, with the training semantic resources.
Further, the system is also used to:By association Cheng Chi at least one described son training parallel training.
The embodiment of the invention also provides a kind of nonvolatile computer storage media, computer storage medium is stored with meter
Calculation machine executable instruction, the computer executable instructions can be performed in above-mentioned any means embodiment and are used for voice dialogue platform
Semantic resources training method;
As an implementation, nonvolatile computer storage media of the invention is stored with the executable finger of computer
It enables, computer executable instructions are set as:
In response to the click to dictionary button is created in dictionary edit page, article editing interface is generated, wherein institute's predicate
Editing interface includes at least entry and adds button;
In response to the click of the publication button of speech production in voice dialogue platform, in the history training of the speech production
The first version number that training script is obtained in information, obtains the second edition number of the training script of the voice dialogue platform;
When the first version number and the second edition number are consistent, unmodified data pair in the speech production is determined
The first semantic resources answered, by corresponding second semantic resources of modification data;
The training classification of second semantic resources is determined by modification data according to described, is second semantic resources
Determine at least one son training;
By at least one described son training, third semantic resources are determined, by first semantic resources and the third
Semantic resources combine, to complete the training of the semantic resources.
As a kind of non-volatile computer readable storage medium storing program for executing, it can be used for storing non-volatile software program, non-volatile
Property computer executable program and module, such as the corresponding program instruction/mould of the method for the test software in the embodiment of the present invention
Block.One or more program instruction is stored in non-volatile computer readable storage medium storing program for executing, when being executed by a processor, is held
The semantic resources training method for voice dialogue platform in the above-mentioned any means embodiment of row.
Non-volatile computer readable storage medium storing program for executing may include storing program area and storage data area, wherein storage journey
It sequence area can application program required for storage program area, at least one function;Storage data area can be stored according to test software
Device use created data etc..In addition, non-volatile computer readable storage medium storing program for executing may include that high speed is deposited at random
Access to memory, can also include nonvolatile memory, a for example, at least disk memory, flush memory device or other are non-
Volatile solid-state part.In some embodiments, it includes relative to place that non-volatile computer readable storage medium storing program for executing is optional
The remotely located memory of device is managed, these remote memories can be by being connected to the network to the device of test software.Above-mentioned network
Example include but is not limited to internet, intranet, local area network, mobile radio communication and combinations thereof.
The embodiment of the present invention also provides a kind of electronic equipment comprising:At least one processor, and with described at least one
The memory of a processor communication connection, wherein the memory is stored with the finger that can be executed by least one described processor
Enable, described instruction executed by least one described processor so that at least one described processor be able to carry out it is of the invention any
The step of semantic resources training method for voice dialogue platform of embodiment.
The client of the embodiment of the present application exists in a variety of forms, including but not limited to:
(1) mobile communication equipment:The characteristics of this kind of equipment is that have mobile communication function, and to provide speech, data
Communication is main target.This Terminal Type includes:Smart phone (such as iPhone), multimedia handset, functional mobile phone and low
Hold mobile phone etc..
(2) super mobile personal computer equipment:This kind of equipment belongs to the scope of personal computer, there is calculating and processing function
Can, generally also have mobile Internet access characteristic.This Terminal Type includes:PDA, MID and UMPC equipment etc., such as iPad.
(3) portable entertainment device:This kind of equipment can show and play multimedia content.Such equipment includes:Audio,
Video player (such as iPod), handheld device, e-book and intelligent toy and portable car-mounted navigation equipment.
(4) other electronic devices having data processing function.
Herein, relational terms such as first and second and the like be used merely to by an entity or operation with it is another
One entity or operation distinguish, and without necessarily requiring or implying between these entities or operation, there are any this reality
Relationship or sequence.Moreover, the terms "include", "comprise", include not only those elements, but also including being not explicitly listed
Other element, or further include for elements inherent to such a process, method, article, or device.Do not limiting more
In the case where system, the element that is limited by sentence " including ... ", it is not excluded that including process, method, the article of the element
Or there is also other identical elements in equipment.
The apparatus embodiments described above are merely exemplary, wherein described, unit can as illustrated by the separation member
It is physically separated with being or may not be, component shown as a unit may or may not be physics list
Member, it can it is in one place, or may be distributed over multiple network units.It can be selected according to the actual needs
In some or all of the modules achieve the purpose of the solution of this embodiment.Those of ordinary skill in the art are not paying creativeness
Labour in the case where, it can understand and implement.
Through the above description of the embodiments, those skilled in the art can be understood that each embodiment can
It realizes by means of software and necessary general hardware platform, naturally it is also possible to pass through hardware.Based on this understanding, on
Stating technical solution, substantially the part that contributes to existing technology can be embodied in the form of software products in other words, should
Computer software product may be stored in a computer readable storage medium, such as ROM/RAM, magnetic disk, CD, including several fingers
It enables and using so that a computer equipment (can be personal computer, server or the network equipment etc.) executes each implementation
Method described in certain parts of example or embodiment.
Finally it should be noted that:The above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations;Although
Present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that:It still may be used
To modify the technical solutions described in the foregoing embodiments or equivalent replacement of some of the technical features;
And these are modified or replaceed, technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution spirit and
Range.