Using Machine Reading Comprehension (MRC), the machine can analyze the semantics of the article and the connection between article and question, and then answer the relevant questions accurately according to the content of the article.
Applications & Benefits
With a large number of text data in various industries, such as listing FAQs, manual data processing has become a bottleneck due to slow processing speed and high cost. The MRC of CITC uses multi-turn dialogue and memory feature fusion technology to extract features of historical data, historical response and current problem, so that can make machine understand the context more accurately. According to the user's description of the problem, machine can find the related content from the documentation, and give a solution to the user.
Through the interactive dialogue, MRC can understand the key content in the dialogue. In addition to replying to key questions, it can also be used to guide shopping. This system can be applied to various AI artificial intelligence dialogue application fields, such as: commodity consultation, virtual customer service, dialogue business, virtual employee assistants, etc.