https://github.com/baidu/AnyQ

FAQ-based Question Answering System
https://github.com/baidu/AnyQ

dialogue dialogue-systems faq faq-question-answering-system faq-system faqbot qa-system question-answering semantic-matching

Last synced: 4 months ago
JSON representation

FAQ-based Question Answering System

• Host: GitHub
• URL: https://github.com/baidu/AnyQ
• Owner: baidu
• License: apache-2.0
• Created: 2018-07-02T03:21:55.000Z (over 6 years ago)
• Default Branch: master
• Last Pushed: 2020-11-28T12:32:49.000Z (about 4 years ago)
• Last Synced: 2024-10-15T19:03:23.695Z (4 months ago)
• Topics: dialogue, dialogue-systems, faq, faq-question-answering-system, faq-system, faqbot, qa-system, question-answering, semantic-matching
• Language: C++
• Homepage:
• Size: 5.41 MB
• Stars: 2,584
• Watchers: 121
• Forks: 672
• Open Issues: 72
• Metadata Files:
  • Readme: README.EN.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# AnyQ

AnyQ(ANswer Your Questions) is a configurable & plugable FAQ-based Question Answering framework. SimNet, a Semantic Matching Framework developed by Baidu-NLP, is also conveyed with AnyQ. 

In our FAQ-based QA framework, which is designed to be configurable and plugable, all the processes or functions are plugins. Developers can easily designed their own processes and add to our framework, so they can quickly build QA system for their own application.  

SimNet, first designed in 2013 by Baidu-NLP, is a flexiable semantic matching framework which is widely used in many applications in Baidu. SimNet consists of the neural network structure BOW、CNN、RNN and MM-DNN. Meanwhile, we have implemented more state-of-the-art structures such as MatchPyramid、MV-LSTM、K-NRM. SimNet has a unified interface, implemented with PaddleFluid and Tensorflow. Models trained using SimNet can be easily added into our AnyQ framework, through which we can augment our semantic matching ability.

The overall framework of AnyQ is as follows:



## Acknowledgments & Statements

This work is supported by the National Key R&D Program of China (No.  **2018YFB1004300** ).