FAQ library matching method and system based on twin network
A technology of twinning network and matching method, which is applied in the field of FAQ question answering library matching based on twinning network, which can solve the problems of poor matching accuracy of questions and so on.
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0069] figure 1 It is a schematic diagram of the steps of a twin network-based FAQ question-answer database matching method provided by the present invention. Such as figure 1 As shown, this embodiment discloses a specific implementation of a twin network-based FAQ question-answer library matching method (hereinafter referred to as "method").
[0070] The method of the present invention has three precondition assumptions: 1, guarantee the timeliness of the data of the FAQ question-and-answer database, that is, include certain hot issues that users often raise; 2, under the premise of keeping accurate, the sentences are as short as possible; The historical record of user questions, sorting the data of the FAQ question and answer database by the number of occurrences.
[0071] Specifically, the method disclosed in this embodiment mainly includes the following steps:
[0072] Step S1: Input two questions into the embedding layer and process them, and input the word vectors and...
Embodiment 2
[0100] In combination with the twin network-based FAQ question-answer database matching method disclosed in Embodiment 1, this embodiment discloses a specific implementation example of a twin network-based FAQ question-answer database matching system (hereinafter referred to as the "system").
[0101] refer to Figure 9 As shown, the system includes:
[0102] Input module 11: input two question sentences in the embedding layer and process them, and input the word vector and word vector fusion in the said question sentence after processing to the encoding layer;
[0103] Information extraction module 12: using twin network architecture in the encoding layer, using two bidirectional LSTM networks, respectively performing context encoding on the fused word vector and the word vector from the two directions of the question sentence, And perform question feature extraction, and input the state value of the hidden layer of each time step of the LSTM network to the attention layer; ...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More