The intelligent Q&A robot--GTCOM's self-developed intelligent Q&A system--can answer the factual, computational and logical reasoning questions that users ask in the form of natural language. The application scenarios include chat robots, intelligent customer-service robots, personal assistant robots and intelligent creation robots. Using a technology roadmap that combines retrieval and generation, it not only finds the most appropriate answers from the domain knowledge base according to the user's questions, doing so on the basis of search-engine technology, but it also acquires a large Q&A corpus through a seq2seq model based on deep-learning technology in order to generate answers according to the questions it receives. GTCOM has conducted in-depth research on the practical issues of contextual information expression and use, safe answers to questions, and information consistency in multiple rounds of dialogues, obtaining impressive results in the process.
