Target-level event text classification method based on topic perception and collaborative thought

• Main Authors: CHENG YINGHAN, SHI CHONGYANG, XIAO LIANG
• Format: Patent
• Language: Chinese; English
• Published: 16.07.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS

The invention relates to a target-level event text classification method based on theme perception and collaborative thought, and belongs to the technical field of artificial intelligence and natural language processing. According to the method, different aspects/implicit topics considered by different categories are obtained by using a topic modeling technology; according to the method, a heterogeneous topic map is constructed, association is captured according to implicit topics of different category labels, cooperation signals among texts, topics and classification labels are fully utilized through a cooperation propagation aggregation module, and direct supervision is applied to the heterogeneous map to learn characterization of the texts, the implicit topics and the category labels; a scoring module is designed, a two-way encoder based on Transform is seamlessly integrated to represent Bert and CPA for reasoning, semantic information provided by the Bert and cooperation information brought by the CPA are