Summary of the invention
The object of the invention is to overcome the deficiencies in the prior art, the invention provides a kind of mutual implementation method and system thereof of distributed intelligence that adopts natural language, can finish challenge, greatly improve the success ratio of finishing the work.
In order to address the above problem, the present invention proposes a kind of mutual implementation method of distributed intelligence that adopts natural language, described method comprises:
The human-computer interactive control module receives the request with the natural language formal representation that the human-computer interaction device sends;
Described human-computer interactive control module sends to natural language parsing/Understanding Module with described request with the natural language formal representation, described natural language parsing/Understanding Module obtains the result that represents in the machine language mode according to described request with the natural language formal representation, and the described result that represents in the machine language mode is returned to described human-computer interactive control module;
Described human-computer interactive control module is converted to the executable order of machine with the described result that represents in the machine language mode, and the executable order of described machine is sent to execution/response module;
Described execution/response module to described machine executable order carry out task analysis, obtain the tasks carrying routing table, and described tasks carrying routing table sent to the fill order distribution module;
Described fill order distribution module is distributed to corresponding at least one processing node on the network according to received execution routing table with task, the return results of finishing the work and processing and obtain to represent with machine language by described at least one processing node, and the described return results that represents with machine language sent to natural language conversion/express module;
Described natural language is changed/is expressed module the described return results that represents with machine language is converted to return results with natural language expressing, and described return results with natural language expressing is sent to described human-computer interactive control module;
Described human-computer interactive control module sends to described human-computer interaction device with described return results with natural language expressing, is shown by described human-computer interaction device.
Preferably, described human-computer interactive control module sends to natural language parsing/Understanding Module with described request with the natural language formal representation, described natural language parsing/Understanding Module is according to the described result that represents in the machine language mode with the request acquisition of natural language formal representation, and the step that the described result that represents in the machine language mode is returned to described human-computer interactive control module comprises:
Described human-computer interactive control module according to the policy control rule to described request with the natural language formal representation judgement of classifying, obtain the route Compass of this processing, according to described route Compass described request with the natural language formal representation is sent to described natural language parsing/Understanding Module;
Described natural language parsing/Understanding Module is processed described request with the natural language formal representation, the result that acquisition represents in the machine language mode, and the described result that represents in the machine language mode sent to described human-computer interactive control module.
Preferably, in described human-computer interactive control module the described result that represents in the machine language mode is converted to the executable order of machine, and the executable order of described machine is sent to before the step of execution/response module, also comprise:
Described human-computer interactive control module is estimated judgement to the described result that represents in the machine language mode, obtains the optimal processing result.
Preferably, in described human-computer interactive control module described return results with natural language expressing is sent to described human-computer interaction device, before the step that is shown by described human-computer interaction device, also comprises:
Described human-computer interactive control module is estimated judgement to described return results with natural language expressing, obtains optimum return results;
Then described human-computer interactive control module sends to described human-computer interaction device with described return results with natural language expressing, and the step that is shown by described human-computer interaction device comprises:
Described human-computer interactive control module sends to described human-computer interaction device with described optimum return results, is shown by described human-computer interaction device.
Preferably, described natural language parsing/Understanding Module comprises a plurality of natural language parsing/Understanding Modules that are distributed on the network, comprises a plurality of parsings/automatic processing module of understanding and a plurality of parsing/understand artificial treatment module.
Preferably, described natural language is changed/is expressed module and comprises that a plurality of natural languages that are distributed on the network change/express module, comprises a plurality of conversions/automatic processing module of expression and a plurality of conversion/expression artificial treatment module.
Preferably, described processing node comprises automatic processing node and artificial treatment node.
Correspondingly, the present invention also provides a kind of mutual realization system of distributed intelligence that adopts natural language, and described system comprises: module, processing node are changed/expressed to human-computer interaction device, human-computer interactive control module, natural language parsing/Understanding Module, execution/response module, fill order distribution module, natural language;
Described human-computer interactive control module is used for receiving the request with the natural language formal representation that described human-computer interaction device sends, and described request with the natural language formal representation is sent to described natural language parsing/Understanding Module;
Described natural language parsing/Understanding Module is used for obtaining the result that represents in the machine language mode according to the request with the natural language formal representation from described human-computer interactive control module of sending, and the described result that represents in the machine language mode is returned to described human-computer interactive control module;
Described human-computer interactive control module also is used for the described result that represents in the machine language mode is converted to the executable order of machine, and the executable order of described machine is sent to described execution/response module;
Described execution/response module is used for to described machine executable order and carries out task analysis, obtains the tasks carrying routing table, and described tasks carrying routing table is sent to described fill order distribution module;
Described fill order distribution module is used for according to received execution routing table task being distributed to corresponding at least one processing node on the network, the return results of finishing the work and processing and obtain to represent with machine language by described at least one processing node, and the described return results that represents with machine language is sent to described natural language change/express module;
Module is changed/expressed to described natural language for the return results that the described return results that represents with machine language is converted to natural language expressing, and described return results with natural language expressing is sent to described human-computer interactive control module;
Described human-computer interactive control module also is used for described return results with natural language expressing is sent to described human-computer interaction device, is shown by described human-computer interaction device.
Preferably, described human-computer interactive control module also is used for according to the policy control rule described request with the natural language formal representation judgement of classifying, obtain the route Compass of this processing, according to described route Compass described request with the natural language formal representation is sent to described natural language parsing/Understanding Module;
Described natural language parsing/Understanding Module also is used for described request with the natural language formal representation is processed, the result that acquisition represents in the machine language mode, and the described result that represents in the machine language mode sent to described human-computer interactive control module.
Preferably, described human-computer interactive control module also is used for the described result that represents in the machine language mode is estimated judgement, obtains the optimal processing result.
Preferably, described human-computer interactive control module also is used for described return results with natural language expressing is estimated judgement, obtain optimum return results, and described optimum return results is sent to described human-computer interaction device, shown by described human-computer interaction device.
Preferably, described processing node comprises automatic processing node and artificial treatment node.
In embodiments of the present invention, the user is by a plurality of natural language parsing/Understanding Modules on the tactful route sensing serial or parallel startup network of human-computer interactive control module (IRCM), natural language is resolved understanding, be converted to machine language, again machine language is distributed to a plurality of processing nodes on the network, a plurality of processing node collaborative works, solve user's problem, traditional point-to-point method of service is changed into any to the method for service of multiple spot, after the inquiry result is returned to the user, can finish challenge, greatly improve the success ratio of finishing the work; What user and human-computer interaction device's input and output were used alternately all is natural languages, and namely machine adapts to people's mode, can give full expression to user's script intention, has reduced the professional knowledge requirement to user (user), very easy to use; The user carries out interactive information by intelligent interaction machine and a plurality of automatic or artificial treatment node that is distributed on the network, has overcome the defective of operator's knowledge and data deficiency, can better meet user's needs; Behind user's off-line, service still can continue, and artificial and automatic processing node also can continue as the answer that the user seeks problem, and the continuous follow-up work cycle task of growing is to solve the problem that complexity and working time grow; Can process better privacy concern, make the user can propose relatively easily the privacy demand, satisfy the demand that the user relates to the privacy aspect.
Embodiment
Below in conjunction with the accompanying drawing in the embodiment of the invention, the technical scheme in the embodiment of the invention is clearly and completely described, obviously, described embodiment only is the present invention's part embodiment, rather than whole embodiment.Based on the embodiment among the present invention, those of ordinary skills belong to the scope of protection of the invention not making the every other embodiment that obtains under the creative work prerequisite.
Fig. 1 is the general principles structural drawing of the mutual realization system of the distributed intelligence of employing natural language of the embodiment of the invention, and as shown in Figure 1, this system can comprise the human-computer interaction device; Human-computer interactive control module (Interactive Response Control Module, IRCM) with human-computer interaction device's connection; Change/express module, execution/response module (Execution Control Module, EXCM) with natural language parsing/Understanding Module, the natural language of IRCM connection respectively; Fill order distribution module and execution result with the EXCM connection returns module respectively; Automatic processing node 1 ... N1, artificial treatment node 1 ... N2 returns module with execution result respectively by Wireless/wired communication network and the fill order distribution module is connected.The human-computer interaction device can be made of terminal keyboard, display, microphone, earphone, loudspeaker, the mutual equipment of human and computer, this human-computer interaction device is including, but not limited to mobile phone, computing machine, personal digital assistant (Personal Digital Assistant, PDA) etc.
Below in conjunction with Fig. 1 the mutual implementation method of distributed intelligence of the employing natural language of the embodiment of the invention is described in detail.
Fig. 2 shows the schematic flow sheet of the mutual implementation method of the distributed intelligence of employing natural language of the embodiment of the invention, and as shown in Figure 2, the method comprises:
S201, the human-computer interactive control module receives the request with the natural language formal representation that the human-computer interaction device sends, and will send to the request of natural language formal representation natural language parsing/Understanding Module;
S202, the result that natural language parsing/Understanding Module represents in the machine language mode according to the request acquisition with the natural language formal representation, and will return to the result that the machine language mode represents the human-computer interactive control module;
S203, the human-computer interactive control module will be converted to the result that the machine language mode represents the executable order of machine, and the executable order of machine is sent to execution/response module;
S204, execution/response module to machine executable order carry out task analysis, obtain the tasks carrying routing table, and the tasks carrying routing table sent to the fill order distribution module;
S205, the fill order distribution module is distributed to corresponding at least one processing node on the network according to received execution routing table with task, the return results of finishing the work and processing and obtain to represent with machine language by at least one processing node, and will send to the return results that machine language represents natural language conversion/express module;
S206, natural language change/express module will be converted to the return results that machine language represents return results with natural language expressing, and will send to the return results of natural language expressing the human-computer interactive control module;
S207, the human-computer interactive control module will send to the human-computer interaction device with the return results of natural language expressing, be shown by the human-computer interaction device.
In implementation, be to receive the user by the human-computer interaction device to be inputted with the request of natural language formal representation.
Further, S202 comprises:
The human-computer interactive control module according to the policy control rule to the judgement of classifying with the request of natural language formal representation, obtain the route Compass of this processing, will send to the request of natural language formal representation natural language parsing/Understanding Module according to the route Compass;
Natural language parsing/Understanding Module is to processing with the request of natural language formal representation, obtains the result that represents in the machine language mode, and will send to the result that the machine language mode represents the human-computer interactive control module.
In implementation, natural language parsing/Understanding Module comprises a plurality of natural language parsing/Understanding Modules that are distributed on the network, by a plurality of natural language parsing/Understanding Module collaborative work on the network, further, natural language parsing/Understanding Module can comprise a plurality of parsings/automatic processing module of understanding and a plurality of parsing/understand artificial treatment module.
In fact, in S202, human-computer interactive control module (IRCM) sends to natural language parsing/Understanding Module with the request of natural language expressing and resolves, understands processing, specifically can be by following process implementation:
1, the request with natural language expressing that will receive of the natural language of human-computer interactive control module (IRCM) transmitting-receiving buffer portion sends parsing/comprehension strategy control module to, parsing/comprehension strategy control module draws the route Compass of this processing according to the judgement of classifying of policy control rule;
2, parsing/comprehension strategy control module sends to automatic parsing/understand routing module control and manually resolve/understand routing module control with the route Compass with the request of natural language expressing;
3, automatically resolve/understand one or more natural languages that routing module control starts on the network simultaneously and resolve/understand automatic processing module, manually resolve/understand one or more natural languages that routing module control starts on the network simultaneously and resolve/understand the work of artificial treatment module cooperative, and reclaim that automatic processing module is resolved/understood to each natural language and the result that the artificial treatment module returns represents in the machine language mode is resolved/understood to natural language, each result is estimated judgement, obtain the optimal processing result, send the optimal processing result to parsing/comprehension strategy control module;
4, parsing/comprehension strategy control module draws final optimal processing result to from automatic parsing/understand routing module control with manually resolve/understand the optimal processing result that routing module control returns and estimate judgement.
In addition, before S207, also comprise: the human-computer interactive control module is estimated judgement to the return results with natural language expressing, obtains optimum return results;
Then S207 is specially: the human-computer interactive control module sends to the human-computer interaction device with optimum return results, is shown by the human-computer interaction device.
In implementation, natural language is changed/is expressed module and comprises that a plurality of natural languages that are distributed on the network change/express module, change/express module cooperative work by a plurality of natural languages on the network, further, natural language is changed/is expressed module and can comprise a plurality of conversions/automatic processing module of expression and a plurality of conversion/expression artificial treatment module.
In fact, in S205 to S207, human-computer interactive control module (IRCM) will send with the return results that machine language represents natural language conversion/express module conversion, express processing to, and detailed process can realize by following steps:
1, the execution result of human-computer interactive control module (IRCM) receives the return results that buffer portion represents the machine language that receives and sends conversion/expression strategy control module to, conversion/expression strategy control module draws the route Compass of this processing according to the judgement of classifying of policy control rule;
2, conversion/expression strategy control module sends to automatic conversion/expression routing module control and artificial conversion/expression routing module control with the route Compass with the return results that machine language represents;
3, automatically conversion/express routing module control starts the automatic processing module of one or more conversions on the network/express simultaneously, artificial conversion/express routing module control starts one or more conversions on the network/expression artificial treatment module cooperative work simultaneously, and reclaim the automatic processing module of each conversion/express and conversion/expression artificial treatment module with the return results of natural language formal representation, each return results is estimated judgement, obtain optimum return results, send optimum return results to conversion/expression strategy control module;
4, conversion/expression strategy control module is estimated judgement to the result who returns from automatic conversion/expression routing module control and artificial conversion/express routing module control, draws final optimum return results.
Fig. 3 is the structural principle schematic diagram of the human-computer interactive control module (IRCM) of the embodiment of the invention, below in conjunction with Fig. 3 the principle of work of the human-computer interactive control module (IRCM) of the embodiment of the invention is further set forth.
As shown in Figure 3, human-computer interactive control module (IRCM) further comprises: natural language transmitting-receiving buffer portion, receive and dispatch the parsing that buffer portion is connected/comprehension strategy control module and conversion/expression strategy control module with natural language; Wherein, parsing/comprehension strategy control module respectively with automatic parsing/understand routing module control, manually resolve/understand routing module control and the order synthesis module is connected; Conversion/expression strategy control module receives buffer portion with execution result and is connected with automatic conversion/expression routing module control, artificial conversion/expression routing module control respectively.
The natural language transmitting-receiving buffer portion of human-computer interaction device and human-computer interactive control module (IRCM) connects, and user's demand is sent to the natural language transmitting-receiving buffer portion of human-computer interactive control module (IRCM) according to the form of natural language text strings or voice string by the human-computer interaction device; Natural language transmitting-receiving buffer portion sends it to parsing/comprehension strategy control module and reads, and by the judgement of classifying of policy control rule, draws the route Compass of this processing; Send to automatic parsing/understand routing module control and manually resolve/understand routing module control with the route Compass and with the request of natural language formal representation; Startup is automatically resolved/is understood routing module control and manually resolves/understand routing module control and carry out work.
Automatically resolve/understand routing module control and manually resolve/understand routing module control and start one or more parsings in one or more parsings of being distributed on the network in K1 the node/automatic processing module of understanding or K2 the node/understand artificial treatment module according to the Way guidance serial or parallel of route Compass, request with natural language expressing is resolved, understand, be converted to machine language, and reclaim each node and return result, the result of respectively returning is estimated judgement to obtain the optimal processing result, send the optimal processing result who returns to parsing/comprehension strategy control module.
When automatic parsing/understand routing module control or manually to resolve/understand result that routing module control returns be successfully the time, parsing/comprehension strategy control module sends result to the order synthesis module, result after will resolving synthesizes the executable order of the machine that represents in the machine language mode by the order synthesis module, and sends execution/response module (EXCM) to.Execution/response module (EXCM) is carried out task analysis to the order that represents in the machine language mode, obtain the tasks carrying routing table, send the tasks carrying routing table to the fill order distribution module, the fill order distribution module is distributed to corresponding one or more automatic processing nodes and artificial treatment node on the network according to the tasks carrying routing table that receives with task order, automatically processing node and artificial treatment node return to execution result with return results and return module, and execution result returns module return results is returned to execution/response module (EXCM); Execution/response module (EXCM) is estimated judgement to return results, obtain optimum return results, receive buffer portion by execution result return results is passed to conversion/expression strategy control module, conversion/expression strategy control module draws the route Compass of this processing by the judgement of classifying of policy control rule; Send to automatic conversion/expression routing module control and artificial conversion/expression routing module control with the route Compass and with the return results that the machine language mode represents; Start automatic conversion the/expression routing module control and artificial conversion the/expression routing module control and carry out work.Automatically conversion/express routing module control and artificial conversion/express routing module control guides serial or parallel to start to be distributed on the network one or more conversions in one or more conversions in K3 the node/automatic processing module of expression and K4 the node/expression artificial treatment module according to the route of route Compass, return results is changed, form with natural language is expressed, and reclaim the return results of each node, each return results is estimated judgement obtain optimum return results, send optimum return results to conversion/expression strategy control module.
The optimum return results that conversion/expression strategy control module will be received sends natural language transmitting-receiving buffer portion to, by natural language transmitting-receiving buffer portion optimum return results is returned to the human-computer interaction device, is shown to the user.
When automatic parsing/understand routing module control with manually resolve/understand result that routing module control returns when being unsuccessfully, parsing/comprehension strategy control module will be resolved failed information exchange and be crossed natural language transmitting-receiving buffer portion and return to the human-computer interaction device and show or express to the user.
Implement embodiment of the method for the present invention, have following useful technique effect:
1, the user is by a plurality of natural language parsing/Understanding Modules on the tactful route sensing serial or parallel startup network of human-computer interactive control module (IRCM), natural language is resolved understanding, be converted to machine language, again machine language is distributed to a plurality of processing nodes on the network, a plurality of processing node collaborative works, solve user's problem, traditional point-to-point method of service is changed into any to the method for service of multiple spot, after the inquiry result is returned to the user, can finish challenge, greatly improve the success ratio of finishing the work;
2, the mutual use of user and human-computer interaction device's input and output all is natural languages, and namely machine adapts to people's mode, can give full expression to user's script intention, has reduced the professional knowledge requirement to user (user), very easy to use;
3, the user carries out interactive information by intelligent interaction machine and a plurality of automatic or artificial treatment node that is distributed on the network, has overcome the defective of operator's knowledge and data deficiency, can better meet user's needs;
4, behind user's off-line, service still can continue, and artificial and automatic processing node also can continue as the answer that the user seeks problem, and the continuous follow-up work cycle task of growing is to solve the problem that complexity and working time grow;
5, can process better privacy concern, make the user can propose relatively easily the privacy demand, satisfy the demand that the user relates to the privacy aspect.
In embodiments of the present invention, in the mutual realization system of the distributed intelligence that adopts natural language, human-computer interactive control module (IRCM) is classified the request with the natural language formal representation that the human-computer interaction device sends by the policy control rule, draw the route Compass that parsing/understanding is processed, start a plurality of parsings/automatic processing node of understanding or parsing/understand artificial treatment node cooperative work, form task that the machine language form represents and process request and pass to tasks carrying processing controls part; Tasks carrying processing controls part is processed according to transmitting the machine language commands task of coming, and the task result is returned to human-computer interactive control module (IRCM) in the machine language mode; Human-computer interactive control module (IRCM) will be classified by the policy control rule with the task result that the machine language mode represents, draw the route Compass that conversion/expression is processed, start a plurality of conversions/automatic processing node of expression or conversion/expression artificial treatment node cooperative work, the task result that represents in the machine language mode is changed, expressed, be converted to the task result of natural language formal representation and pass to the human-computer interactive control module and be shown to the user.
The embodiment of the invention also provides a kind of mutual realization system of distributed intelligence that adopts natural language, as shown in Figure 4, this realization system comprises: module 45, processing node 46 are changed/expressed to human-computer interaction device 40, human-computer interactive control module 41, natural language parsing/Understanding Module 42, execution/response module 43, fill order distribution module 44, natural language; Wherein,
Human-computer interactive control module 41 is used for receiving the request with the natural language formal representation that human-computer interaction device 40 sends, and will send to the request of natural language formal representation natural language parsing/Understanding Module 42;
Natural language parsing/Understanding Module 42 is used for obtaining the result that represents in the machine language mode according to the request with the natural language formal representation from human-computer interactive control module 41 of sending, and will return to the result that the machine language mode represents human-computer interactive control module 41;
Human-computer interactive control module 41 also is used for being converted to the result that the machine language mode represents the executable order of machine, and the executable order of machine is sent to execution/response module 43;
Execution/response module 43 is used for to machine executable order and carries out task analysis, obtains the tasks carrying routing table, and the tasks carrying routing table is sent to fill order distribution module 44;
Fill order distribution module 44 is used for according to received execution routing table task being distributed to corresponding at least one processing node 46 on the network, the return results of finishing the work and processing and obtain to represent with machine language by at least one processing node, and will send to the return results that machine language represents natural language conversion/express module 45;
Module 45 is changed/expressed to natural language for the return results that will be converted to the return results that machine language represents with natural language expressing, and will send to the return results of natural language expressing human-computer interactive control module 41;
Human-computer interactive control module 41 also is used for and will sends to human-computer interaction device 40 with the return results of natural language expressing, is shown by human-computer interaction device 40.
In the implementation, human-computer interactive control module 41 also is used for according to the policy control rule the judgement of classifying with the request of natural language formal representation, obtain the route Compass of this processing, will send to the request of natural language formal representation natural language parsing/Understanding Module 42 according to the route Compass;
Natural language parsing/Understanding Module 42 also is used for processing with the request of natural language formal representation, obtains the result that represents in the machine language mode, and will send to the result that the machine language mode represents human-computer interactive control module 41.
Human-computer interactive control module 41 also is used for the result that represents in the machine language mode is estimated judgement, obtains the optimal processing result.
Human-computer interactive control module 41 also is used for the return results with natural language expressing is estimated judgement, obtains optimum return results, and optimum return results is sent to human-computer interaction device 40, is shown by human-computer interaction device 40.
Processing node 46 specifically comprises automatic processing node and artificial treatment node.
Below in conjunction with the principle of work of the implementation method of the embodiment of the invention and Fig. 1, Fig. 3 the function of each functional module of the realization system of the embodiment of the invention is further described in detail.
The human-computer interaction device 40: the user instruction with the natural language input that will receive passes to human-computer interactive control module (IRCM) 41; Receive the result with natural language expressing that man-machine interaction control module (IRCM) 41 is returned;
Execution/response module (EXCM) 43: the order that represents in the machine language mode that receives that man-machine interaction control module (IRCM) 41 sends forms the tasks carrying routing table after analyzing, and passes to fill order distribution module 44; Control fill order distribution module 44 is distributed to the automatic processing node of processing node 46(or artificial treatment node); The return results that represents in the machine language mode that receives that execution result returns that module returns is estimated judgement to it, forms optimum return results and passes to human-computer interactive control module (IRCM) 41;
Fill order distribution module 44: receive machine language commands and tasks carrying routing table that execution/response module (EXCM) 43 sends, it is distributed to corresponding automatically processing node and artificial treatment node;
Execution result returns module: receive each automatic processing node and artificial treatment node and return result and send execution/response module (EXCM) 43 to what the machine language form represented;
Automatically processing node: the tasks carrying order of fill order distribution module 44 distributions processed and the task result is returned to execution result return module, it comprises a plurality of automatic processing nodes that are distributed on the network;
The artificial treatment node: the tasks carrying order of fill order distribution module 44 distribution processed and the task result is returned to execution result return module, it comprises a plurality of artificial treatment nodes that are distributed on the network;
Human-computer interactive control module (IRCM) 41: control receives natural literal or the language strings that human-computer interaction device 40 sends, and sends natural language parsing/Understanding Module 42 to; Control receives the result that machine language form that natural language parsing/Understanding Module 42 sends represents, sends execution/response module (EXCM) 43 to after synthesizing order; The return results that represents in the machine language mode that control receives that execution/response module (EXCM) 43 sends sends natural language conversion/express module 45 to; Control receives the nature task action result that language conversion/expression module 45 is passed back with the natural language expressing form, and sends human-computer interaction device 40 to;
Natural language parsing/Understanding Module 42: the natural language text strings that human-computer interactive control module (IRCM) 41 is sent or language strings are resolved and are understood, and natural language text strings or language conversion are become and can and return to human-computer interactive control module (IRCM) 41 by the machine language of machine recognition and execution;
Module 45 is changed/expressed to natural language: the task action result of returning in the machine language mode that human-computer interactive control module (IRCM) 41 sends is changed, expressed, be converted to the expression-form of natural language, and return to human-computer interactive control module (IRCM) 41.
In the implementation, human-computer interactive control module (IRCM) 41 further comprises:
Parsing/comprehension strategy control module: the request of natural language being received and dispatched the natural language expressing that buffer portion sends draws the route Compass of this processing by the judgement of classifying of policy control rule; Send to automatic parsing/understand routing module control and manually resolve/understand routing module control with the route Compass with the request of natural language expressing; Startup is automatically resolved/is understood routing module control and manually resolves/understand routing module control and carry out work, result to successfully resolved is analyzed judgement, send final optimum return results to the order synthesis module in the machine language mode, return to human-computer interaction device in the natural language mode by natural language transmitting-receiving buffer portion with resolving failed result, demonstration or expression are to the user;
Automatically resolve/understand routing module control: receive route Compass that parsings/comprehension strategy control module sends, with the request of natural language formal representation and the instruction of startup work, start the automatic processing module of one or more parsings of being distributed on the network/understand according to the Way guidance serial or parallel of route Compass, and reclaim each working node and return result, the result of respectively returning is estimated judgement form the optimal processing result, send the optimal processing result who returns to parsing/comprehension strategy control module;
Manually resolve/understand routing module control: receive route Compass that parsings/comprehension strategy control module sends, with the request of natural language expressing and the instruction of startup work, start one or more parsings of being distributed on the network/understand artificial treatment module according to the Way guidance serial or parallel of route Compass, and reclaim the return results of each working node, each return results is estimated judgement form optimum return results, send optimum return results to parsing/comprehension strategy control module;
Order synthesis module: receive parsing that parsing/comprehension strategy control module sends, understand the result, and it is synthesized the executable machine language commands of machine send execution/response module EXCM to;
Natural language transmitting-receiving buffer portion: receive the request with the natural language formal representation that the human-computer interaction device sends, and send it to parsing/comprehension strategy control module and process; The result of the parsing failure that reception parsing/comprehension strategy control module is returned receives conversion/expression strategy control module with the return results of natural language expressing, and sends return results to human-computer interaction device to be shown to the user;
Conversion/expression strategy control module: the execution response results that represents in the machine language mode that execution result is received that buffer portion sends by the judgement of classifying of policy control rule, draws the route Compass of this processing; Send to automatic conversion/expression routing module control and artificial conversion/expression routing module control with the route Compass with the result that machine language is expressed; Start automatic conversion the/expression routing module control and artificial conversion the/expression routing module control and carry out work, and reclaim the return results of each working node, return results with natural language expressing is analyzed judgement, form final optimum return results and send natural language transmitting-receiving buffer portion to;
Automatically conversion/express routing module control: the result that receive route Compass that conversion/expression strategy control module transmits, represents with the machine language form and the instruction of startup work; The automatic processing module of one or more conversions of guiding serial or parallel to start to be distributed on the network according to the route of route Compass/express, and reclaim the return results of each working node, each return results is estimated judgement form optimum return results, send optimum return results to conversion/expression strategy control module;
Artificial conversion/express routing module control: the result that receive route Compass that conversion/expression strategy control module transmits, represents in the machine language mode and the instruction of startup work; Start one or more conversions of being distributed on the network/expression artificial treatment module according to the Way guidance serial or parallel of route Compass, and reclaim the return results of each working node, each return results is estimated judgement form optimum return results, send optimum return results to conversion/expression strategy control module;
Execution result receives buffer portion: receive the return results that execution/response module (EXCM) represents in the machine language mode and send conversion/expression strategy control module to.
Principle of work part for each functional module in the mutual realization system of the distributed intelligence of the employing natural language of the embodiment of the invention, interactive mode between function realization, principle of work and the functional module of realization system and functional module thereof is identical with the description of the method for the embodiment of the invention, and other function that relates to functional module is here set forth no longer one by one.
Implement in the system embodiment of the present invention, have following useful technique effect:
1, the user is by a plurality of natural language parsing/Understanding Modules on the tactful route sensing serial or parallel startup network of human-computer interactive control module (IRCM), natural language is resolved understanding, be converted to machine language, again machine language is distributed to a plurality of processing nodes on the network, a plurality of processing node collaborative works, solve user's problem, traditional point-to-point method of service is changed into any to the method for service of multiple spot, after the inquiry result is returned to the user, can finish challenge, greatly improve the success ratio of finishing the work;
2, the mutual use of user and human-computer interaction device's input and output all is natural languages, and namely machine adapts to people's mode, can give full expression to user's script intention, has reduced the professional knowledge requirement to user (user), very easy to use;
3, the user carries out interactive information by intelligent interaction machine and a plurality of automatic or artificial treatment node that is distributed on the network, has overcome the defective of operator's knowledge and data deficiency, can better meet user's needs;
4, behind user's off-line, service still can continue, and artificial and automatic processing node also can continue as the answer that the user seeks problem, and the continuous follow-up work cycle task of growing is to solve the problem that complexity and working time grow;
5, can process better privacy concern, make the user can propose relatively easily the privacy demand, satisfy the demand that the user relates to the privacy aspect.
By embodiment described above the present invention has been described; it will be appreciated that; above disclosed only is a kind of preferred embodiment of the present invention; certainly can not limit with this interest field of the present invention; for those skilled in the art; under the prerequisite that does not break away from the principle of the invention, can also make some improvements and modifications, these improvements and modifications also are considered as protection scope of the present invention.