Semantic decision method, system and medium based on whitelist

Abstract

The present invention provides a semantic decision method, system and medium based on a whitelist, and relates to the technical field of human-computer dialogue. The method comprises: step S1: constructing a non-basic domain semantic understanding microservice and a basic domain semantic understanding microservice according to business needs; step S2: constructing a whitelist according to business logic; step S3: inputting text into the non-basic domain semantic understanding microservice and the basic domain semantic understanding microservice at the same time to obtain corresponding results; step S4: judging whether the non-basic domain skills are in the whitelist according to the results, and whether the confidence of the non-basic domain skills meets the set value or not, and obtaining the final semantic result. The present invention can improve the accuracy and performance of semantic understanding by taking the whitelist decision as the final semantic result, while ensuring the user experience of human-computer dialogue.

Description

Technical Field

[0001] The present invention relates to the technical field of human-computer dialogue, and in particular to a whitelist-based semantic decision method, system and medium. Background Art

[0002] Natural language is the crystallization of human wisdom. People have long expected to interact with computers in the form of natural language. With the development of natural language processing technology and the rise of deep learning methods, human-computer dialogue systems have become a new research hotspot. Semantic understanding is a very important part of human-computer dialogue, especially in in-vehicle voice interaction systems. The semantic understanding model of some domain skills requires more than one million user corpora to train and learn in order to achieve good results. Due to the lack of user corpora, external basic domain semantic understanding services may be used first. At this time, the problem of making decisions based on non-basic domain semantic understanding results and basic domain semantic understanding results will arise. If the decision is not good, it may affect the accuracy and performance of the final semantic results and cannot guarantee a good user experience of human-computer dialogue.

[0003] The invention patent with publication number CN104008169B discloses a semantic-based geographic annotation content security inspection method. It uses key technologies such as keyword matching, feature extraction, semantic similarity calculation and rule matching to build a geographic information annotation content security classification system, a banned word library, a white list library and a map review rule library. It realizes a geographic annotation content security inspection method that integrates banned word matching, automatic annotation classification, white list filtering and map public representation rule inspection.

[0004] The existing technology compares the confidence of semantic understanding in the basic domain with the confidence of semantic understanding in the non-basic domain, and takes the larger one as the best semantic result. However, this may not necessarily select the better semantic result, and the business skills in different domains are not very comparable. Summary of the invention

[0005] In view of the defects in the prior art, the present invention provides a whitelist-based semantic decision method, system and medium.

[0006] According to a whitelist-based semantic decision method, system and medium provided by the present invention, the scheme is as follows:

[0007] In a first aspect, a whitelist-based semantic decision method is provided, the method comprising:

[0008] Step S1: construct non-basic domain semantic understanding microservices and basic domain semantic understanding microservices according to business requirements;

[0009] Step S2: construct a whitelist according to business logic;

[0010] Step S3: Input the text into the non-basic domain semantic understanding microservice and the basic domain semantic understanding microservice at the same time to obtain corresponding results;

[0011] Step S4: Determine whether the non-basic domain skill is in the whitelist based on the result, and whether the confidence of the non-basic domain skill is above the set value, and obtain the final semantic result.

[0012] Preferably, step S3 includes: inputting the text into the non-basic domain semantic understanding microservice and the basic domain semantic understanding microservice at the same time, first obtaining the non-basic domain semantic result, and then obtaining the basic domain semantic understanding result;

[0013] The non-basic domain semantic understanding service accesses the intranet, while the basic domain semantic understanding service accesses the public network.

[0014] Preferably, step S4 includes: judging whether the non-basic domain skill is in the whitelist according to the non-basic domain semantic result obtained first, and the confidence of the non-basic domain skill is above 0.9; if so, the non-basic domain semantic result is the final semantic result, otherwise the basic domain semantic result is the final semantic result.

[0015] In a second aspect, a whitelist-based semantic decision system is provided, the system comprising:

[0016] Module M1: Construct non-basic domain semantic understanding microservices and basic domain semantic understanding microservices according to business needs;

[0017] Module M2: construct a whitelist based on business logic;

[0018] Module M3: Input the text into the non-basic domain semantic understanding microservice and the basic domain semantic understanding microservice at the same time to obtain the corresponding results;

[0019] Module M4: Determine whether the non-basic domain skills are in the whitelist based on the results, and whether the confidence of the non-basic domain skills is above the set value, and obtain the final semantic result.

[0020] Preferably, the module M3 comprises: inputting the text into the non-basic domain semantic understanding microservice and the basic domain semantic understanding microservice at the same time, obtaining the non-basic domain semantic result first, and then obtaining the basic domain semantic understanding result;

[0021] The non-basic domain semantic understanding service accesses the intranet, while the basic domain semantic understanding service accesses the public network.

[0022] Preferably, the module M4 includes: judging whether the non-basic domain skill is in the whitelist according to the non-basic domain semantic result obtained first, and the confidence of the non-basic domain skill is above 0.9; if so, the non-basic domain semantic result is the final semantic result, otherwise the basic domain semantic result is the final semantic result.

[0023] Compared with the prior art, the present invention has the following beneficial effects:

[0024] 1. The invention uses a whitelist and single-domain confidence comprehensive strategy to make semantic result decisions on the results of semantic understanding of the basic domain and non-basic domain, thereby improving the performance and accuracy of semantic understanding while also ensuring the user experience of in-vehicle voice dialogue;

[0025] 2. The present invention microservices the semantic understanding of business skills in basic domains and non-basic domains respectively, which can solve the problems of large memory usage when loading different business skill semantic understanding models in basic domains and non-basic domains, incompatibility between different versions of the same installation package, and disadvantages for service engineering. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:

[0027] Figure 1 It is a schematic diagram of the overall process of the present invention. DETAILED DESCRIPTION

[0028] The present invention is described in detail below in conjunction with specific embodiments. The following embodiments will help those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those of ordinary skill in the art, several changes and improvements can also be made without departing from the concept of the present invention. These all belong to the protection scope of the present invention.

[0029] The embodiment of the present invention provides a semantic decision method based on a whitelist, wherein the basic domain skills and non-basic domain skills in the present invention are divided according to the business skill fields of the in-vehicle voice dialogue system in the industry, and the business skill fields that are standard or basically available in the in-vehicle voice dialogue system in the industry are the basic fields. For example: business skills such as navigation, weather and music. As long as there is an in-vehicle voice dialogue system in the industry, basically there will be these business skills. Non-basic domain skills are not standard in-vehicle voice dialogue systems in the industry, and not all in-vehicle voice dialogue systems have them. In order to differentiate from everyone's in-vehicle voice dialogue system and highlight the characteristics and personalization of their own in-vehicle voice dialogue system, some business skills are defined by themselves. For example: business skills such as novice sailing.

[0030] The whitelist is a list of business skills. For example, the whitelist includes basic domain skills: navigation, weather, music and chatting, and non-basic domain skills: beginner’s guide.

[0031] Both the basic domain semantic understanding microservice and the non-basic domain semantic understanding microservice perform semantic understanding on user corpus text, but the focus is different. The basic domain semantic understanding microservice mainly performs semantic understanding on the basic domain user corpus text, while the non-basic domain semantic understanding microservice mainly performs semantic understanding on the non-basic domain user corpus text. The input of the basic domain and non-basic domain semantic understanding microservices are both user corpus texts, and the output is the semantic result of the user corpus text. For example, the basic domain has weather, navigation, music and small talk skills, and the non-basic domain has novice sailing and small talk skills. For example: the input user corpus text is: How is the weather today? The output of the non-basic domain semantic understanding microservice is small talk skills, and the output of the basic domain semantic understanding microservice is weather query skills; the final semantic result is the semantic result of the semantic understanding of the basic domain business skills; for example: the input user corpus text is: How to use the in-vehicle voice dialogue system? The output of the non-basic domain semantic understanding microservice is the novice sailing skills, and the output of the basic domain semantic understanding microservice is small talk skills. At this time, the final semantic result is the semantic result of the semantic understanding of the non-basic domain business skills.

[0032] For the two semantic understanding results of the basic domain and non-basic domain semantic understanding microservices, the non-basic domain semantic understanding microservice returns results faster than the basic domain semantic understanding microservice, because the basic domain business skill semantic understanding microservice is an external network, while the non-basic domain business skill semantic understanding microservice is an internal network. Therefore, although the user corpus text enters the basic domain and non-basic domain business skill semantic understanding microservices at the same time, the basic domain business skill semantic understanding microservice returns semantic results faster, so the final semantic result decision should be made in two steps instead of one step. If it is made in one step, it will affect the performance of obtaining the best semantic result. You can first look at the semantic result of the basic domain business skill semantic understanding microservice. If the business skill is in the whitelist and the confidence is relatively high, you can directly use the semantic result of the non-basic domain business skill semantic understanding as the best semantic result without waiting. If the semantic result of the basic domain business skill semantic understanding microservice is not in the whitelist or is in the whitelist but the confidence is not high, you need to wait for the semantic result of the basic domain business skill semantic understanding and use the semantic result as the best semantic result.

[0033] Microservices: Docker architecture is mainly used to make the environment and functions independent so that multiple functions do not affect each other.

[0034] Specifically, refer to Figure 1 As shown, the present invention provides a whitelist-based semantic decision method, which specifically includes the following steps:

[0035] Step S1: Construct non-basic domain semantic understanding microservices and basic domain semantic understanding microservices according to business requirements.

[0036] Step S2: Construct a whitelist according to business logic.

[0037] Step S3: Input the text into the non-basic domain semantic understanding microservice and the basic domain semantic understanding microservice at the same time to obtain corresponding results;

[0038] The specific step is as follows: the text is input into the non-basic domain semantic understanding microservice and the basic domain semantic understanding microservice at the same time, and the non-basic domain semantic result is obtained first, and then the basic domain semantic understanding result is obtained, because the non-basic domain semantic understanding service accesses the intranet, and the basic domain semantic understanding service accesses the public network.

[0039] Step S4: Determine whether the non-basic domain skill is in the whitelist based on the result, and whether the confidence of the non-basic domain skill is above the set value, and obtain the final semantic result.

[0040] Specifically, it is determined whether the non-basic domain skill is in the whitelist according to the non-basic domain semantic result obtained first, and the confidence of the non-basic domain skill is above 0.9. If it is satisfied, the non-basic domain semantic result is the final semantic result, otherwise the basic domain semantic result is the final semantic result.

[0041] The present invention also provides a whitelist-based semantic decision system, the system comprising:

[0042] Module M1: Construct non-basic domain semantic understanding microservices and basic domain semantic understanding microservices according to business needs;

[0043] Module M2: construct a whitelist based on business logic;

[0044] Module M3: Input the text into the non-basic domain semantic understanding microservice and the basic domain semantic understanding microservice at the same time to obtain the corresponding results;

[0045] Module M4: Determine whether the non-basic domain skills are in the whitelist based on the results, and whether the confidence of the non-basic domain skills is above the set value, and obtain the final semantic result.

[0046] In module M3: the text is input into the non-basic domain semantic understanding microservice and the basic domain semantic understanding microservice at the same time, and the non-basic domain semantic result is obtained first, and then the basic domain semantic understanding result is obtained;

[0047] The non-basic domain semantic understanding service accesses the intranet, while the basic domain semantic understanding service accesses the public network.

[0048] In module M4: determine whether the non-basic domain skill is in the whitelist based on the non-basic domain semantic result obtained first, and the confidence of the non-basic domain skill is above 0.9. If it is satisfied, the non-basic domain semantic result is the final semantic result, otherwise the basic domain semantic result is the final semantic result.

[0049] The embodiments of the present invention provide a whitelist-based semantic decision method, system and medium, which use a whitelist to decide the final semantic result so as to improve the accuracy and performance of semantic understanding while ensuring the user experience of human-computer dialogue.

[0050] Those skilled in the art know that, in addition to realizing the system and its various devices, modules, and units provided by the present invention in a purely computer-readable program code, it is entirely possible to realize the same functions in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, the system and its various devices, modules, and units provided by the present invention can be considered as a hardware component, and the devices, modules, and units included therein for realizing various functions can also be regarded as structures within the hardware component; the devices, modules, and units for realizing various functions can also be regarded as both software modules for realizing the method and structures within the hardware component.

[0051] The above describes the specific embodiments of the present invention. It should be understood that the present invention is not limited to the above specific embodiments, and those skilled in the art can make various changes or modifications within the scope of the claims, which does not affect the essence of the present invention. In the absence of conflict, the embodiments of the present application and the features in the embodiments can be combined with each other arbitrarily.

Claims

1. A semantic decision method based on whitelist, It is characterized in that include: Step S1: construct non-basic domain semantic understanding microservices and basic domain semantic understanding microservices according to business requirements; Step S2: construct a whitelist according to business logic; Step S3: Input the text into the non-basic domain semantic understanding microservice and the basic domain semantic understanding microservice at the same time to obtain corresponding results; Step S4: judging whether the non-basic domain skill is in the whitelist according to the result, and whether the confidence of the non-basic domain skill is above the set value, and obtaining the final semantic result; The step S3 includes: inputting the text into the non-basic domain semantic understanding microservice and the basic domain semantic understanding microservice at the same time, first obtaining the non-basic domain semantic result, and then obtaining the basic domain semantic understanding result; The non-basic domain semantic understanding service accesses the intranet, while the basic domain semantic understanding service accesses the public network; The step S4 includes: judging whether the non-basic domain skill is in the whitelist according to the non-basic domain semantic result obtained first, and the confidence of the non-basic domain skill is above 0.9; if so, the non-basic domain semantic result is the final semantic result, otherwise the basic domain semantic result is the final semantic result.

2. A semantic decision system based on whitelist, It is characterized in that include: Module M1: Construct non-basic domain semantic understanding microservices and basic domain semantic understanding microservices according to business needs; Module M2: construct a whitelist based on business logic; Module M3: Input the text into the non-basic domain semantic understanding microservice and the basic domain semantic understanding microservice at the same time to obtain the corresponding results; Module M4: Determine whether the non-basic domain skills are in the whitelist based on the results, and whether the confidence of the non-basic domain skills is above the set value, and obtain the final semantic result; The module M3 includes: inputting text into a non-basic domain semantic understanding microservice and a basic domain semantic understanding microservice at the same time, first obtaining a non-basic domain semantic result, and then obtaining a basic domain semantic understanding result; The non-basic domain semantic understanding service accesses the intranet, while the basic domain semantic understanding service accesses the public network; The module M4 includes: judging whether the non-basic domain skill is in the whitelist according to the non-basic domain semantic result obtained first, and the confidence of the non-basic domain skill is above 0.9; if so, the non-basic domain semantic result is the final semantic result, otherwise the basic domain semantic result is the final semantic result.

3. A computer-readable storage medium storing a computer program, It is characterized in that When the computer program is executed by a processor, the steps of the method according to claim 1 are implemented.

Images