Method and apparatus for outputting information

An apparatus for outputting information, comprising:

• at least one processor;
• and
• a memory storing instructions, wherein the instructions when executed by the at least one processor, cause the at least one processor to perform operations, the operations comprising:
• acquiring a to be identified sentence;
• performing word segment on the to be identified sentence to obtain a word sequence;
• and
• inputting the word sequence into a pre-trained multi-task element recognition model based on sequence labeling and entity word predicting, and outputting identified entity words, entity categories, and entity word positions, wherein the multi-task element recognition model comprises a sequence labeling network for performing a sequence labeling task and an entity word predicting network for performing an entity word predicting task, the sequence labeling network is fused with the entity word predicting network through a fusion module.

A non-transitory computer readable medium, storing a computer program thereon, wherein the program, when executed by a processor, causes the processor to perform operations, the operations comprising:

• acquiring a to be identified sentence;
• performing word segment on the to be identified sentence to obtain a word sequence;
• and
• inputting the word sequence into a pre-trained multi-task element recognition model based on sequence labeling and entity word predicting, and outputting identified entity words, entity categories, and entity word positions, wherein the multi-task element recognition model comprises a sequence labeling network for performing a sequence labeling task and an entity word predicting network for performing an entity word predicting task, the sequence lab ng network is fused with the entity word predicting network through a fusion module.

The non-transitory computer readable medium of claim wherein the acquiring a training sample set comprises:

• acquiring an unstructured/semi-structured document;
• performing sentence segment on the document;
• and
• for each clause, performing word segmentation on the clause to obtain a word sequence, and perform entity labeling on the clause by a rule matching method to obtain an entity word, an entity category and an entity word position of the clause as labeling information.