FIELD: information technology.
SUBSTANCE: invention relates to recognition of complete sentences for a human-machine intelligent question-answer system. Method of recognizing complete sentences for a human-machine intelligent inquiry-response system includes: receiving current sentence input by user; input of current sentence into pre-trained model of recognition of segmented sentences in order to obtain first probability of current offer being a complete sentence, wherein the segmented suggestions recognition model is used to determine the first probability in accordance with the probability corresponding to the current language Chinese model and the probability corresponding to the language model based on the recurrent neural network; and determining the current sentence as a complete sentence if the first probability is greater than a given first threshold.
EFFECT: high accuracy of recognizing complete sentences for a human-machine intelligent question-answer system.
12 cl, 6 dwg
| Title | Year | Author | Number |
|---|---|---|---|
| METHOD AND APPARATUS FOR ANALYZING SEMANTIC INFORMATION | 2017 |
|
RU2704531C1 |
| METHOD AND DEVICE FOR SENDING AN ELECTRONIC BOOK | 2016 |
|
RU2704495C1 |
| METHOD OF CREATING MODEL FOR ANALYSING DIALOGUES BASED ON ARTIFICIAL INTELLIGENCE FOR PROCESSING USER REQUESTS AND SYSTEM USING SUCH MODEL | 2019 |
|
RU2730449C2 |
| METHOD AND EQUIPMENT FOR RECOGNIZING EMOTIONS IN SPEECH | 2019 |
|
RU2720359C1 |
| METHOD FOR COMPRESSION OF NEURAL NETWORK MODEL AND METHOD AND APPARATUS FOR LANGUAGE CORPORA TRANSLATION | 2019 |
|
RU2749970C1 |
| METHOD AND APPARATUS FOR VOICE ACTIVATION | 2019 |
|
RU2760368C1 |
| METHOD, APPARATUS AND DEVICE FOR NATURAL LANGUAGE PROCESSING | 2019 |
|
RU2726739C1 |
| BAYESIAN RAREFACTION OF RECURRENT NEURAL NETWORKS | 2018 |
|
RU2702978C1 |
| MACHINE TRAINING | 2005 |
|
RU2391791C2 |
| METHOD OF MULTIMODAL CONTACTLESS CONTROL OF MOBILE INFORMATION ROBOT | 2020 |
|
RU2737231C1 |
