Short text entity disambiguation method based on multi-task learning

• Main Authors: XIAO JIANMAO, LEI GANG, YIN ZIYU, CAO YUANLONG, YI YUGEN, WANG YONGDI
• Format: Patent
• Language: Chinese; English
• Published: 20.09.2022
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS

The invention provides a short text entity disambiguation method based on multi-task learning, which belongs to the technical field of natural language processing, and comprises the following steps: respectively carrying out multi-task learning and constructing a short text entity disambiguation model based on a knowledge enhanced pre-training language model; wherein the multiple tasks comprise a reference and entity semantic similarity disambiguation task, a reference mask and entity semantic similarity disambiguation task and a reference classification task; and obtaining the sum of semantic similarity scores of the candidate entities on the reference and entity semantic similarity disambiguation task and the reference mask and entity semantic similarity disambiguation task, and taking the candidate entities with high semantic similarity scores as prediction entities for finally completing entity disambiguation. Through a multi-task learning mode, the pre-training model makes full use of referred context in