Abstract: A corpus annotation apparatus obtains a corpus set provided by a user through a client, where the corpus set includes a plurality of semantic categories of corpuses that the user expects to annotate, determines a manual annotation corpus and an automatic annotation corpus falling within a target semantic category in the corpus set, obtains a manual annotation result of the manual annotation corpus, and automatically annotates the automatic annotation corpus based on the manual annotation result of the manual annotation corpus. The manual annotation result and an automatic annotation result that correspond to the automatic annotation corpus are used as training data to train an inference model.

Abstract: A modal information completion method, an apparatus, and a device are provided. A completion apparatus first obtains a modal information group, wherein the modal information group includes at least two pieces of modal information. Then, the completion apparatus may determine, based on an attribute of the modal information group, whether a part or all of first modal information in the modal information group is missing. Subsequently, the completion apparatus determines a target feature vector of the first modal information based on a preset feature vector mapping relationship and a feature vector of second modal information in the modal information group, so that accuracy of the target feature vector of the first modal information is ensured.

Abstract: A question answering method based on a knowledge graph are disclosed, including: After obtaining a first natural query statement including a first query entity and first entity characteristic information, the terminal queries the knowledge graph based on the first query entity and the first entity characteristic information, to obtain a first result. The terminal obtains a second natural query statement after outputting the first result. The terminal determines a missing part in the second natural query statement based on the first query entity or the first entity characteristic information. The terminal queries the knowledge graph based on the missing part in the second natural query statement to obtain a second result, and outputs the second result.