Yitu Technology invents intelligent interactive solution for epidemic information

The COVID-19 pandemic swept the world in 2020, causing immeasurable impact on the global economy and ecology. In order to enable people to better understand epidemic consultation and related knowledge, many online platforms have launched related content query functions.

At present, based on the use of some traditional technologies for information processing of major infectious disease epidemics, such as COVID-19, plague, highly pathogenic influenza, Ebola and other major infectious disease epidemics, the following problems often arise in the implementation of prevention and control work:

1) When using traditional search engines to learn about epidemic information, due to the simple search rules, the returned results cannot accurately match the problem, and the massive search results will cause confusion for users and make it difficult to make a decision;

2) When using intelligent voice robots to learn about epidemic information, they can only solve questions raised by users that have clear answers, and cannot connect corresponding product functions, so the problems they can solve are limited.

Therefore, in order to meet the public's understanding of epidemic prevention knowledge and solve the problems in the existing technology, Yitu Technology applied for an invention patent entitled "Intelligent interaction method, device and storage medium for epidemic situation" on September 24, 2020 (application number: 202011015236.4), and the applicant is Shanghai Yitu Network Technology Co., Ltd.

Based on the relevant information currently disclosed in the patent, let us take a look at this intelligent interactive solution for the epidemic.

As shown in the figure above, it is a flow chart of the intelligent interaction method for epidemics invented in this patent. First, the system responds to the user's first interaction operation and calls the epidemic knowledge base from the database. The first interaction operation is, for example, the user's voice or text input information. Secondly, the system continues to respond to the user's second interaction operation and recognizes the user's intention based on the second operation.

The first interaction is that the user can call the epidemic knowledge base of a certain type of epidemic through human-computer interaction operations. For example, if the user inputs an operation containing "new coronavirus pneumonia", the system can call the epidemic knowledge base about new coronavirus pneumonia based on the "new coronavirus pneumonia" contained in the operation; the second interactive operation is further operation performed by the user, such as asking questions such as "What is the new coronavirus?", so that the system can understand the user's specific intentions and provide answers to the user.

Next, the system will detect whether the user's intention matches the content in the current corpus, and then determine the target content in the corpus through the guidance of other content in the corpus that is semantically related to the content. The purpose of this step is to guide the user through the two questions with the highest similarity to the current question based on the similarity between the questions in the corpus and the questions asked by the user, so as to gradually provide the user with accurate answers through similar questions in the corpus.

Finally, the system can output the content in the epidemic knowledge base corresponding to the target content to the user, thereby answering the user's questions and relevant questions.

As shown in the figure above, this is the interaction process between the user and the intelligent system shown in the patent. It can be seen that the interaction process is carried out in the form of a question and an answer. The user asks "What is the new coronavirus" through the second interactive operation, and the corresponding response output is "XXXXXX..."

For example, the user proposes "self-check for avian influenza" through the second interactive operation. Usually, the process of screening for avian influenza requires information such as body temperature and epidemiological survey. The solution in this patent can use these contents as elements to generate replies, and provide interactive replies such as body temperature, whether to sneeze, and options related to epidemiological survey for users to interact with.

The above is the intelligent interactive solution for the epidemic invented by Yitu Technology. This solution is based on AI and big data technology, which enables the results returned by the search engine when understanding epidemic information to accurately match the user's intentions, and provide targeted and accurate interactive content in massive search results, thereby significantly improving the efficiency of searching for the epidemic and related knowledge.