Semantic analysis method and device

Abstract

The embodiment of the invention discloses a semantic analysis method and device. A specific embodiment of the method comprises the steps of obtaining an offline vertical class set; in response to a received to-be-identified statement, detecting a current network state; for each vertical class in the vertical class set, carrying out the offline semantic analysis of a heuristic derivation mode and the offline semantic analysis of a sample generalization model mode on the to-be-identified statement according to the vertical class to respectively obtain a first analysis result and a second analysis result of the vertical class, if the first analysis result meets a preset condition, taking the first analysis result as an offline candidate analysis result of the vertical class, otherwise, carrying out offline fusion on the first analysis result and the second analysis result to generate an offline candidate analysis result of the vertical class; and if the current network state is network disconnection or network instability, selecting a final analysis result from the offline candidate analysis results of different vertical classes according to the priorities and / or historical dialogue information of the vertical classes. According to the embodiment, the accuracy and flexibility of the semantic analysis are improved.

Description

technical field [0001] The embodiments of the present disclosure relate to the field of computer technology, and in particular, to a semantic analysis method and device. Background technique [0002] With the rapid development and popularization of AIoT (AI+Internet of Things), more and more smart devices have the ability of voice interaction, such as smart vehicles, smart wearable devices, smart speakers, etc. The intelligent interaction capability of smart devices depends on intelligent services in the cloud, including voice wake-up, voice recognition, semantic analysis, speech synthesis, etc. Among them, voice wake-up, voice recognition, and synthesis all have mature offline technical solutions. Only the semantic analysis capability has no mature feasible method, and can only rely entirely on cloud services. In smart vehicle scenarios, cars may enter tunnels or remote areas. When the network is unstable or disconnected, online semantic analysis cannot serve normally. The...

AI Technical Summary

Problems solved by technology

[0004] 1. Inflexible, poor generalization

A large number of rules and corpus need to be manually configured, which cannot cope with the diversity of dialogue;

[0005] 2. Poor maintainability

When there are too many rules, the maintenance cost will become higher and higher, and the rule base cannot be managed effectively;

[0006] 3. Poor performance

As the number increases, the efficiency of regular expressions and text matching will be significantly reduced, bringing performance pressure to smart devices;

[0007] 4. Unable to reuse online analysis service

The implementation methods of offline analysis capability and online analysis service are different, so that the model and data of online analysis service cannot be reused by offline analysis capability. Offline analysis capability needs to be redeveloped, and the effect is quite different from that of online analysis service;

[0008] 5. Poor scalability and high migration costs

There is no mature technical framework, self-contained system, unable to quickly expand to new fields, unable to migrate offline analysis capabilities to other intelligent platforms

Images