CN115146615A - Natural language processing method, system, equipment and readable storage medium - Google Patents
Natural language processing method, system, equipment and readable storage medium Download PDFInfo
- Publication number
- CN115146615A CN115146615A CN202211068041.5A CN202211068041A CN115146615A CN 115146615 A CN115146615 A CN 115146615A CN 202211068041 A CN202211068041 A CN 202211068041A CN 115146615 A CN115146615 A CN 115146615A
- Authority
- CN
- China
- Prior art keywords
- arbitration
- natural language
- data
- rule
- language processing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/205—Parsing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/445—Program loading or initiating
- G06F9/44521—Dynamic linking or loading; Link editing at or after load time, e.g. Java class loading
Abstract
The invention discloses a natural language processing method and a system, comprising the following steps: acquiring data processed by a natural language processing model, and performing pre-arbitration on the data; and inputting the data passing through the pre-arbitration into a rule engine, and outputting a final result. According to the invention, data are processed based on a plurality of natural language processing models, a plurality of arbitration and rule engines are introduced, the arbitration result is input into the rule engine, the final result is obtained, the problems of low development efficiency and high iteration cost can be effectively solved, and the experience diversity and intelligence of the system are improved.
Description
Technical Field
The invention relates to the technical field of natural language processing, in particular to a natural language processing method, a natural language processing system, natural language processing equipment and a readable storage medium in the fields of Internet and artificial intelligence.
Background
Conventional dialog systems at present tend to build a dialog experience based on a single model when landing on a real project. Due to the limitations of the AI algorithm itself and the ever-increasing complexity of user experience in the actual business scenario, a single model cannot cover a complete experience design. To remedy this drawback, recent partial new projects attempt to adopt a multi-model approach, i.e., building dialog systems based on multiple NLP models. When a plurality of models coexist, due to the difference of algorithms in the models, the output skill confidence statistic dimensionality, the output field and the output intention division dimensionality are not uniform, and the result of which model is selected cannot be directly judged based on the output skill confidence statistic dimensionality, the output field and the output intention division dimensionality. And in order to guarantee the consistency and continuity of user experience in the whole conversation life cycle, the reliability requirement on the trust obtaining result is higher than that of a single model. Thus, the complexity of multimodal dialog state management is far greater than that of a single model. Due to the influence of business factors, enterprise competition and the like, the current project adopting the scheme tends to put the arbitration pressure on the client side, and the client side carries out customized development according to different business scenes and combination fields and intentions.
Therefore, regardless of a single model or multiple models, the dialog systems in different projects are very coupled to the service design of a single project, so that the system cannot be rapidly migrated to other projects, the development cost cannot be effectively reduced, and the development efficiency cannot be improved. Meanwhile, since the judgment logic is coupled in each client, any arbitration rule modification initiated by factors such as iterative experience must be firstly upgraded. When the client is a vehicle end or the like, the cost of one-time over-the-air downloading is very high, the upgrading period is long, and a user cannot obtain better experience in time.
Disclosure of Invention
The invention mainly aims to provide a natural language processing method, a system, equipment and a readable storage medium aiming at the defects of low development efficiency and high iteration cost caused by non-uniform division dimensions when multiple models coexist in the prior art.
In order to achieve the above object, the present invention provides a natural language processing method, comprising the steps of:
acquiring data processed by a natural language processing model, and performing pre-arbitration on the data;
and inputting the data passing through the pre-arbitration into a rule engine, and outputting a final result.
In the natural language processing method provided by the invention, the steps of acquiring data processed by a natural language processing model and pre-arbitrating the data comprise:
acquiring natural language understanding data processed by at least one natural language processing model;
performing pre-arbitration on the natural language understanding data according to the natural language understanding data processed by each natural language processing model;
and carrying out secondary pre-arbitration according to the dialogue state management confidence coefficient generated after the processing of each natural language processing model.
In the natural language processing method provided by the invention, the arbitration rule at least comprises a rule factor and a rule logic,
loading an arbitration rule defined in advance, and performing weighted sequencing on the result of secondary pre-arbitration, wherein the step comprises the following steps of:
loading the rule factors and the rule logic to a rule engine;
and performing weighted sorting on the results of the secondary pre-arbitration according to the rule factors and the rule logic.
In the natural language processing method provided by the present invention, after the steps of obtaining data processed by a natural language processing model and performing pre-arbitration on the data, the method further comprises:
and if the natural language understanding data does not pass the pre-arbitration or no effective data is returned, directly judging that the conversation is rejected.
In addition, in order to achieve the above object, the present invention further provides a natural language processing system, which comprises an obtaining module, a pre-arbitration module and a processing module,
the acquisition module is used for acquiring data processed by the natural language processing model;
the pre-arbitration module is used for pre-arbitrating data;
the processing module is used for inputting the data passing through the pre-arbitration into the rule engine and outputting a final result.
In the natural language processing system provided by the invention, the pre-arbitration module is further used for pre-arbitrating the natural language understanding data according to the natural language understanding data processed by each natural language processing model, and performing secondary pre-arbitration according to the dialogue state management confidence coefficient generated after the natural language processing model is processed;
the processing module is also used for loading an arbitration rule defined in advance, performing weighted sorting on the result of the secondary pre-arbitration, inputting the data subjected to weighted sorting into the rule engine, and outputting the final result.
In the natural language processing system provided by the invention, the processing module is further used for loading the rule factors and the rule logic in the arbitration rules to the rule engine and performing weighted sequencing on the result of the secondary pre-arbitration according to the rule factors and the rule logic.
In addition, to achieve the above object, the present invention further provides a terminal device, including:
a memory for storing a computer program;
a processor for implementing the steps of the above natural language processing method when executing the computer program.
Further, to achieve the above object, the present invention also provides a computer-readable storage medium storing a computer program which, when executed by a processor, implements the steps of the natural language processing method as above.
The natural language processing method and system of the invention comprises the following steps: acquiring data processed by a natural language processing model, and performing pre-arbitration on the data; and inputting the data passing through the pre-arbitration into a rule engine, and outputting a final result. Data are processed based on a plurality of natural language processing models, multiple times of arbitration and rule engines are introduced, arbitration results are input into the rule engines, final results are obtained, the problems of low development efficiency and high iteration cost can be effectively solved, and the experience diversity and intelligence of the system are improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the provided drawings without creative efforts:
fig. 1 is a flowchart illustrating a natural language processing method according to an embodiment of the present invention.
Fig. 2 is an interaction diagram of pre-arbitration according to an embodiment of the present invention.
FIG. 3 is a diagram illustrating a rule logic design according to an embodiment of the present invention.
Fig. 4 is an interaction diagram of a rule engine according to an embodiment of the present invention.
Fig. 5 is an interaction diagram of a dialog system according to an embodiment of the present invention.
Fig. 6 is a schematic structural diagram of a natural language processing system according to an embodiment of the present invention.
Fig. 7 is a schematic diagram of a terminal device according to an embodiment of the present invention.
Detailed Description
To facilitate an understanding of the invention, the invention will now be described more fully with reference to the accompanying drawings. Exemplary embodiments of the invention are shown in the drawings. This invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. The terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.
In order to better understand the technical solutions, the technical solutions will be described in detail below with reference to the drawings and the specific embodiments of the specification, and it should be understood that the embodiments and specific features of the embodiments of the present invention are detailed descriptions of the technical solutions of the present application, and are not limited to the technical solutions of the present application, and the technical features of the embodiments and examples of the present invention may be combined with each other without conflict.
Referring to fig. 1, fig. 1 is a flowchart illustrating a natural language processing method according to an embodiment of the present invention, in which the natural language processing method includes:
and step S10, acquiring the data processed by the natural language processing model and carrying out pre-arbitration on the data.
The method comprises the steps of obtaining data processed by a natural language processing model and carrying out pre-arbitration on the data, and comprises the following steps: acquiring natural language understanding data processed by at least one natural language processing model, carrying out pre-arbitration on the natural language understanding data according to the natural language understanding data processed by each natural language processing model, managing confidence degrees according to conversation states generated after processing by each natural language processing model, and carrying out secondary pre-arbitration.
The following is a description of an embodiment of the dialogue system.
In one embodiment, a plurality of NLU (natural language understanding) data processed by an NLP (natural language processing) model are acquired, the NLU data includes a field, an intention, and the like, specifically, dialog information is imported according to the requirements of each NLP, and after a result is acquired, the results of the field, the intention, a confidence coefficient, and the like in the result are analyzed and used for arbitration of subsequent steps.
In another embodiment, pre-arbitration is performed based on the NLU results produced by each NLP model. Confidence is the basic arbitration criterion: results below the belonging service confidence threshold are discarded and no longer participate in the subsequent flow.
For example, the DM (session management) part of a certain service is processed in stages and individually, and here, the NLU result passing the screening can be continuously sent to the corresponding DM service to obtain the DM result.
For another example, the DM result of a certain service is output together with the NLU, where the DM result is retained without additional processing for use in the subsequent stage.
There may be a mixture of the two cases, as long as there is a result of NLU, DM when entering the final arbitration by the confidence filtering.
If all NLU results do not pass the pre-arbitration or no valid result is returned, the dialog is directly judged to be rejected here.
And performing secondary pre-arbitration according to the DM confidence coefficient output by each NLP model.
In yet another embodiment, a second pre-arbitration is performed based on the DM confidence produced by each NLP model.
Specifically, the confidence level is a basic arbitration criterion: results below the belonging service confidence threshold are discarded and no longer participate in the subsequent flow.
If all DM results fail pre-arbitration, or no valid results are returned, then the dialog is determined to be rejected directly here.
Referring to fig. 2, when NLU and DM are separately output, it is determined whether NLU data satisfies a pre-arbitration criterion, if not, the dialog is rejected, if the pre-arbitration criterion is satisfied, a DM confidence is obtained, it is determined again whether the pre-arbitration criterion is satisfied, if the pre-arbitration criterion is not satisfied, the dialog is rejected, and if the pre-arbitration criterion is satisfied, the dialog is input to the rules engine.
When NLU and DM are output in combination, NLU data and DM confidence are respectively obtained, whether the NLU data and DM confidence meet the pre-arbitration standard is judged, if the NLU data and DM confidence do not meet the pre-arbitration standard, the dialogue is rejected, and if the NLU data and DM meet the pre-arbitration standard, the dialogue is input into a rule engine.
And step S20, inputting the data passing through the pre-arbitration into a rule engine, and outputting a final result.
Inputting the data passing through the pre-arbitration into a rule engine, and outputting a final result, wherein the step comprises the following steps: and loading an arbitration rule defined in advance, carrying out weighted sorting on the results of secondary pre-arbitration, inputting the data subjected to weighted sorting into a rule engine, and outputting the final result.
The arbitration rules include at least a rule factor and a rule logic.
The method comprises the steps of loading an arbitration rule defined in advance and carrying out weighted sequencing on the result of secondary pre-arbitration, and comprises the following steps: and loading the rule factors and the rule logic to a rule engine, and performing weighted sequencing on the results of the secondary pre-arbitration according to the rule factors and the rule logic.
After the steps of obtaining the data processed by the natural language processing model and performing pre-arbitration on the data, the method further comprises the following steps: and if the natural language understanding data does not pass the pre-arbitration or no effective data is returned, directly judging that the conversation is rejected.
In one embodiment, arbitration rules defined in advance are loaded for weighted ordering.
The arbitration rules may be designed based on a plurality of fact dimensions, split into a plurality of conditional dimensions, each dimension containing a different decision factor (currently input fact object), for example, skill information: domain, intent, etc., current client state (in-vehicle domain for example): front and background App states, device positions and the like, and the current conversation state: single-turn conversations, multiple-turn conversations, etc.
Generally, one condition dimension is selected as an entry dimension, and the decision factors of the other condition dimensions are different in rules of different entry dimensions, and usually a single round or multiple rounds are used as the entry dimensions. This is not essential.
And designing a rule factor and a rule logic of the arbitration part based on the dimensions of multiple conditions and by combining service weighting, and loading the rule factor and the rule logic to a rule engine.
The rule factors are designed to be one-to-one or one-to-many mapping relations based on each decision factor in the multi-condition dimension.
The design of rule logic is based on experience design, the experience design is refined into a clear conversation process in each scene, and then the rule logic is deduced: if the product design needs to block chats after entering multiple rounds of conversations, the design of the rule logic can refer to fig. 3.
The rule logic is essentially a decision stream, so when the decision factor is converted into a rule factor, the data requirements of each actually selected rule engine on the input of the decision stream should be met.
The rule module of a general rule engine supports two functions: the LHS (Left Hand Side) section is conditional branch logic, and the RHS (Right Hand Side) section is execution logic. Therefore, in the rule design, the existing decision factor output by the NLP model is removed, and additional decision factor obtaining actions can be designed at different rule nodes, such as: when the dialogue triggers the holiday question-answer skill rules in the single-turn dialogue scene, whether a business party has some customized operation tasks or not can be inquired additionally, and the part of business specifies which NLP model is born.
In another embodiment, the data that will pass through the pre-arbitration is input to the rules engine, and the final result is output.
Referring to fig. 4, a user customizes an arbitration rule, the arbitration rule is imported into a rule engine, and a result processed by a plurality of NLP models is also imported into the rule engine, so as to obtain a result object.
Referring to fig. 5, a dialog system is taken as an example, a request of dialog information is distributed to each NLP service, when NLU and DM of NLP service 1 are separately output, it is determined whether NLU data meets the requirement of confidence, if not, the dialog is rejected, if meeting the requirement of confidence is met, the confidence of DM is obtained, it is determined again whether meeting the requirement of confidence is met, if not, the dialog is rejected, if meeting the requirement of confidence is met, the dialog is input to the rule engine.
When NLU and DM of NLP service 2 are output in a combined mode, NLU data and DM confidence coefficient are obtained respectively, whether the requirement of meeting the confidence coefficient is met or not is judged, if the requirement of meeting the confidence coefficient is not met, the dialogue is rejected, and if the requirement of meeting the confidence coefficient is met, the dialogue is input into a rule engine.
The user carries out version management and dependency management in the configuration center, the pull rule carries out rule scheduling and mode execution in the rule engine, whether an available result exists or not is judged, if the available result does not exist, the conversation is rejected, and if the available result exists, a result object is output.
Based on the mode, data are processed based on a plurality of natural language processing models, multiple arbitration and rule engines are introduced, arbitration results are input into the rule engines, final results are obtained, the problems of low development efficiency and high iteration cost can be effectively solved, and experience diversity and intelligence of the system are improved.
Referring to fig. 6, correspondingly, the present invention further provides a natural language processing system, an obtaining module 201, a pre-arbitration module 202 and a processing module 203,
the obtaining module 201 is configured to obtain data processed by a natural language processing model;
the pre-arbitration module 202 is used for pre-arbitrating data;
the processing module 203 is used for inputting the data passing through the pre-arbitration into the rule engine and outputting the final result.
Further, in an embodiment of the present invention, the pre-arbitration module 202 is further configured to pre-arbitrate the natural language understanding data according to the natural language understanding data processed by each natural language processing model, and perform secondary pre-arbitration according to a dialog state management confidence generated after processing by each natural language processing model; the processing module 203 is further configured to load an arbitration rule defined in advance, perform weighted sorting on the results of the secondary pre-arbitration, input the data after weighted sorting into the rule engine, and output a final result.
Further, in an embodiment of the present invention, the processing module 203 is further configured to load a rule factor and a rule logic in the arbitration rule to the rule engine, and perform weighted ordering on the result of the secondary pre-arbitration according to the rule factor and the rule logic.
Referring to fig. 7, an embodiment of the present invention further provides a terminal device 100, which may include:
a memory 101 for storing a computer program;
the processor 102, when executing the computer program stored in the memory 101, may implement the following steps:
acquiring data processed by a natural language processing model, and performing pre-arbitration on the data; and inputting the data passing through the pre-arbitration into a rule engine, and outputting a final result.
The embodiment of the invention also provides a computer readable storage medium, the computer readable storage medium stores a computer program, and the computer program can realize the following steps when being executed by a processor;
acquiring data processed by a natural language processing model, and performing pre-arbitration on the data; and inputting the data passing through the pre-arbitration into a rule engine, and outputting a final result.
The computer-readable storage medium may include: a U-disk, a removable hard disk, a Read-Only Memory (ROM) > Random Access Memory (RAM), a magnetic disk, or an optical disk, and other various media capable of storing program codes.
In the description provided herein, numerous specific details are set forth. However, it is understood that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures and techniques have not been shown in detail in order not to obscure an understanding of this description.
Similarly, it should be appreciated that in the foregoing description of exemplary embodiments of the invention, various features of the invention are sometimes grouped together in a single embodiment, figure, or description thereof for the purpose of streamlining the disclosure and aiding in the understanding of one or more of the various inventive aspects. However, the disclosed method should not be interpreted as reflecting an intention that: that the invention as claimed requires more features than are expressly recited in each claim. Rather, as the following claims reflect, inventive aspects lie in less than all features of a single foregoing disclosed embodiment. Thus, the claims following the detailed description are hereby expressly incorporated into this detailed description, with each claim standing on its own as a separate embodiment of this invention.
Those skilled in the art will appreciate that the modules in the device in an embodiment may be adaptively changed and disposed in one or more devices different from the embodiment. The modules or units or components of the embodiments may be combined into one module or unit or component, and furthermore they may be divided into a plurality of sub-modules or sub-units or sub-components. All of the features disclosed in this specification (including any accompanying claims, abstract and drawings), and all of the processes or elements of any method or apparatus so disclosed, may be combined in any combination, except combinations where at least some of such features and/or processes or elements are mutually exclusive. Each feature disclosed in this specification (including any accompanying claims, abstract and drawings) may be replaced by alternative features serving the same, equivalent or similar purpose, unless expressly stated otherwise.
Furthermore, those skilled in the art will appreciate that while some embodiments herein include some features included in other embodiments, rather than other features, combinations of features of different embodiments are meant to be within the scope of the invention and form different embodiments. For example, in the following claims, any of the claimed embodiments may be used in any combination.
Various component embodiments of the invention may be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. Those skilled in the art will appreciate that a microprocessor or Digital Signal Processor (DSP) may be used in practice to implement some or all of the functionality of some or all of the components in accordance with embodiments of the present invention. The present invention may also be embodied as apparatus or device programs (e.g., computer programs and computer program products) for performing a portion or all of the methods described herein. Such programs implementing the present invention may be stored on a computer readable medium or may be in the form of one or more signals. Such a signal may be downloaded from an internet website or provided on a carrier signal or in any other form.
It should be noted that the above-mentioned embodiments illustrate rather than limit the invention, and that those skilled in the art will be able to design alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word "comprising" does not exclude the presence of elements or steps not listed in a claim. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The invention may be implemented by means of hardware comprising several distinct elements, and by means of a suitably programmed computer. In the unit claims enumerating several means, several of these means may be embodied by one and the same item of hardware. The usage of the words first, second and third, etcetera do not indicate any ordering. These words may be interpreted as names.
Claims (10)
1. A natural language processing method, characterized in that the method comprises the steps of:
acquiring data processed by a natural language processing model, and performing pre-arbitration on the data;
and inputting the data passing through the pre-arbitration into a rule engine, and outputting a final result.
2. The natural language processing method of claim 1, wherein the step of obtaining the data processed by the natural language processing model and pre-arbitrating the data comprises:
acquiring natural language understanding data processed by at least one natural language processing model;
performing pre-arbitration on the natural language understanding data according to the natural language understanding data processed by each natural language processing model;
and carrying out secondary pre-arbitration according to the dialogue state management confidence coefficient generated after the processing of each natural language processing model.
3. The natural language processing method of claim 2, wherein the step of inputting the data through the pre-arbitration into the rule engine and outputting the final result comprises:
loading an arbitration rule defined in advance, and performing weighted sequencing on the result of the secondary pre-arbitration;
and inputting the weighted and sequenced data into a rule engine, and outputting a final result.
4. The natural language processing method of claim 3 wherein the arbitration rules include at least a rule factor and a rule logic,
the step of loading the arbitration rules defined in advance and carrying out weighted sequencing on the results of the secondary pre-arbitration comprises the following steps:
loading the rule factor and the rule logic to a rule engine;
and carrying out weighted sequencing on the result of the secondary pre-arbitration according to the rule factor and the rule logic.
5. The natural language processing method of claim 1, wherein after the step of obtaining the data processed by the natural language processing model and pre-arbitrating the data, further comprising:
and if the natural language understanding data does not pass through the pre-arbitration or no effective data is returned, directly judging that the conversation is rejected.
6. A natural language processing system is characterized in that the system comprises an acquisition module, a pre-arbitration module and a processing module,
the acquisition module is used for acquiring data processed by the natural language processing model;
the pre-arbitration module is used for pre-arbitrating the data;
the processing module is used for inputting the data passing through the pre-arbitration into the rule engine and outputting a final result.
7. The natural language processing system of claim 6, wherein the pre-arbitration module is further configured to pre-arbitrate the natural language understanding data according to the natural language understanding data processed by each natural language processing model, and perform secondary pre-arbitration according to a dialog state management confidence generated after the processing of each natural language processing model;
the processing module is further used for loading an arbitration rule defined in advance, performing weighted sorting on the results of the secondary pre-arbitration, inputting the data subjected to weighted sorting into a rule engine, and outputting a final result.
8. The natural language processing system of claim 7 wherein the processing module is further configured to load a rule factor and rule logic in an arbitration rule into a rules engine and to weight-order the results of the secondary pre-arbitration based on the rule factor and the rule logic.
9. A terminal device, characterized in that the terminal device comprises: memory, a processor and a data processing program stored on the memory and executable on the processor, the data processing program, when executed by the processor, implementing the method of any one of claims 1 to 5.
10. A computer-readable storage medium, characterized in that a data processing program is stored on the computer-readable storage medium, which data processing program, when executed by a processor, carries out the method of any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211068041.5A CN115146615A (en) | 2022-09-02 | 2022-09-02 | Natural language processing method, system, equipment and readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211068041.5A CN115146615A (en) | 2022-09-02 | 2022-09-02 | Natural language processing method, system, equipment and readable storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115146615A true CN115146615A (en) | 2022-10-04 |
Family
ID=83416130
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211068041.5A Pending CN115146615A (en) | 2022-09-02 | 2022-09-02 | Natural language processing method, system, equipment and readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115146615A (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105551494A (en) * | 2015-12-11 | 2016-05-04 | 奇瑞汽车股份有限公司 | Mobile phone interconnection-based vehicle-mounted speech recognition system and recognition method |
| CN109949817A (en) * | 2019-02-19 | 2019-06-28 | 一汽-大众汽车有限公司 | Voice referee method and device based on the double speech recognition engines of dual operating systems |
| CN110001549A (en) * | 2019-04-17 | 2019-07-12 | 百度在线网络技术（北京）有限公司 | Method for controlling a vehicle and device |
| CN112820295A (en) * | 2020-12-29 | 2021-05-18 | 华人运通(上海)云计算科技有限公司 | Voice processing device and system, cloud server and vehicle |
| CN113127161A (en) * | 2019-12-30 | 2021-07-16 | 上海仙豆智能机器人有限公司 | Processing engine selection method, processing engine selection device and computer-readable storage medium |
| CN114600099A (en) * | 2019-10-18 | 2022-06-07 | 脸谱科技有限责任公司 | Speech recognition accuracy enhancement using a natural language understanding-based meta-speech system of an assistant system |
| CN114724564A (en) * | 2020-12-18 | 2022-07-08 | 阿里巴巴集团控股有限公司 | Voice processing method, device and system |
-
2022
- 2022-09-02 CN CN202211068041.5A patent/CN115146615A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105551494A (en) * | 2015-12-11 | 2016-05-04 | 奇瑞汽车股份有限公司 | Mobile phone interconnection-based vehicle-mounted speech recognition system and recognition method |
| CN109949817A (en) * | 2019-02-19 | 2019-06-28 | 一汽-大众汽车有限公司 | Voice referee method and device based on the double speech recognition engines of dual operating systems |
| CN110001549A (en) * | 2019-04-17 | 2019-07-12 | 百度在线网络技术（北京）有限公司 | Method for controlling a vehicle and device |
| CN114600099A (en) * | 2019-10-18 | 2022-06-07 | 脸谱科技有限责任公司 | Speech recognition accuracy enhancement using a natural language understanding-based meta-speech system of an assistant system |
| CN113127161A (en) * | 2019-12-30 | 2021-07-16 | 上海仙豆智能机器人有限公司 | Processing engine selection method, processing engine selection device and computer-readable storage medium |
| CN114724564A (en) * | 2020-12-18 | 2022-07-08 | 阿里巴巴集团控股有限公司 | Voice processing method, device and system |
| CN112820295A (en) * | 2020-12-29 | 2021-05-18 | 华人运通(上海)云计算科技有限公司 | Voice processing device and system, cloud server and vehicle |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109583746B (en) | Method and device for setting routing rule of flow and readable storage medium | |
| US20130055211A1 (en) | Client-side policy enforcement of developer api use | |
| CN109561322A (en) | A kind of method, apparatus, equipment and the storage medium of video audit | |
| CN110704594A (en) | Task type dialogue interaction processing method and device based on artificial intelligence | |
| CN110554958A (en) | Graph database testing method, system, device and storage medium | |
| CN103729384A (en) | Information filtering method, system and device | |
| CN106874089A (en) | A kind of processing method of application program self-starting, device and mobile terminal | |
| CN112799782A (en) | Model generation system, method, electronic device, and storage medium | |
| CN114647790A (en) | Big data mining method and cloud AI (Artificial Intelligence) service system applied to behavior intention analysis | |
| CN111950579A (en) | Training method and training device for classification model | |
| CN115146615A (en) | Natural language processing method, system, equipment and readable storage medium | |
| CN115473822B (en) | 5G intelligent gateway data transmission method, system and cloud platform | |
| CN112802478B (en) | Hybrid cloud fusion scheduling method and device | |
| CN115755867A (en) | Vehicle diagnosis method and related device | |
| CN115834375A (en) | Vehicle-side service communication configuration method, system, equipment and storage medium | |
| CN108228292A (en) | A kind of generation method and device at ExtJS interfaces | |
| CN110880099B (en) | Workflow engine-based task event processing method, device, equipment and medium | |
| CN114330675A (en) | Chip, accelerator card, electronic equipment and data processing method | |
| CN113220869A (en) | Text processing device and method for bank peripheral system | |
| CN112948251A (en) | Automatic software testing method and device | |
| CN113297289A (en) | Method and device for extracting business data from database and electronic equipment | |
| CN112084300A (en) | Response information output method and device, electronic equipment and readable storage medium | |
| Brau et al. | A contract-based approach to support goal-driven analysis | |
| CN115993989A (en) | Code identification method, apparatus, computing device, and computer-readable storage medium | |
| Yan et al. | A Test Case Design Method Based on Path Depth Coverage |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20221004 |